Introduction

Fish is widely consumed by humans for its high content in protein and omega-3 polyunsaturated fatty acids, which help abate cholesterol levels and consequently the risk of certain types of cancer and cardiovascular diseases (Storelli 2008). However, teleost fishes are known to accumulate heavy metals (hereafter, ‘metals’ including metalloids: Duffus 2002) in their body by direct uptake from water (via gills) and through diet. Although there can be considerable variability in bioaccumulation amongst fish species depending on life history (Short et al. 2008), the levels of certain metals in fish tissues generally reflect those found in their biotic and abiotic environments (Birge et al. 2000; Wang and Rainbow 2008). Not surprisingly, a high content of heavy metals (mainly, potentially toxic ones such as As, Cd, Cr, Hg, Ni, Pb, but also essential ones such as Cu, Fe, Mn, Se, Zn, if present at elevated concentrations) may lead to severe threat to fish. This, in turn, can represent a serious health hazard for human consumption (Sorensen 1991; Tüzen 2003; Blanco-Penedo et al. 2006).

The common carp Cyprinus carpio is the most widely distributed freshwater fish worldwide (Vilizzi 2012). It represents a major resource for aquaculture production in several countries (http://www.fao.org/fishery/culturedspecies/Cyprinus_carpio/en), makes for valuable and productive fisheries (e.g. Shumka et al. 2008; Harlioğlu 2011) and is sometimes appreciated as a traditional ethnic dish (Balon 1974). In Turkey, C. carpio is native to the ecoregions (sensu Abell et al. 2008) of Thrace, northern Anatolia, western Transcaucasia and upper Tigris and Euphrates (Memiş and Kohlmann 2006; Vilizzi 2012). However, following translocations since the 1960s (Innal and Erk’akan 2006; Çetinkaya 2010; Tarkan et al. 2015), C. carpio is now widespread throughout the rest of the region, where it represents the most important species for inland fisheries (Turkish Statistical Institute 2014). Indeed, stocking of C. carpio by local government agencies especially into newly established (artificial) reservoirs (e.g. Özuluğ et al. 2005; Balık and Ustaoğlu 2006; Önsoy et al. 2011) is common practice, and also provides for an alternative source of revenue for local communities to compensate for the economic losses resulting from private land encroachment following reservoir construction (Gaygusuz et al. 2015; Tarkan et al. 2015).

Traditionally, in Turkey C. carpio has represented a favourite dish especially in the regions of central and eastern Anatolia, where it makes for a large part of the sales of the locally marketed fish (e.g. Orsay and Duman 2008a). To improve the quality of the product for human consumption, studies have therefore investigated the physicochemical and microbiological characteristics of smoked (mirror) C. carpio fillets (Patir and Duman 2006; Duman and Dartay 2007; Duman and Patir 2007; Can et al. 2011), as well as the processing of its flesh into meat balls and sausages (Yanar and Fenercioğlu 1999; Arslan et al. 2001; Can and Coban 2012).

Clearly, given the importance of C. carpio as a food resource in Turkey, it is imperative that a deeper understanding should be gained on the safety of this fish as a consumer’s product, including knowledge of those water bodies sustaining profitable fisheries but more prone to pollution-induced hazards. The objective of the present study was to provide a synthesis on bioaccumulation of metals in C. carpio from water bodies of the region. Based on an extensive review of the literature: (i) a quantitative evaluation was made of the concentration of various metals in body tissues of C. carpio and relative, whenever possible, to established maximum allowed limits for human consumption; (ii) water bodies of the region with above-limit concentration of metals in C. carpio tissues were identified and flagged according to their level of contamination; and (iii) a comparison was made with bioaccumulation of metals in C. carpio tissues from water bodies worldwide with the aim to cast the outcomes of the present study within a global perspective. It is recommended that measures be implemented for enhancing the safety of C. carpio fisheries of the region for human food consumption.

Material and methods

Data collation

Bioaccumulation data for C. carpio in Anatolia were obtained from peer-reviewed papers, thesis dissertations and, occasionally, conference proceedings. A necessary criterion for inclusion of a study into the review was that it should deal with C. carpio sampled under natural conditions. For this reason, laboratory-based (bioassay) studies carried out in situ (Kargın and Erdem 1991; Canlı and Kargın 1995; Barlas 1999b; Cicik 2003; Karaytug et al. 2007; Çoban et al. 2013; Cogun and Kargın 2013; Yeşilbudak and Erdem 2013, 2014) or on fish sampled from one or more water bodies under monitoring (Ozkurt 2000; Sahan et al. 2010; Uysal et al. 2010; Güngördü et al. 2012) were not included. Notably, the dataset provided by Küçükbay and Örün (2003) for Karakaya Reservoir was excluded because of the unrealistic concentration values provided; whereas the study by Çiçek and Koparal (2001) did not present data in a format suitable for analysis and that by Yaman et al. (2013), albeit reporting average concentration values for some metals, relied on samples of C. carpio pooled across several water bodies.

The same criterion used for selection of the Turkish-based studies was applied to bioaccumulation data for C. carpio worldwide. In this case, the studies by Yousafzai et al. (2012) for Kabul River, Pakistan, and by Gummadavelli et al. (2013) for Edulabad Reservoir, India, were excluded from analysis due to unrealistically high values reported. Similarly, the study by Zubcov et al. (2012) from Dubasari Reservoir in Moldova was excluded because of metal concentration values provided only for gonads (i.e. not muscle, liver and/or gill).

Data preparation

For each study from Anatolia, information was retrieved about the water body under investigation, which was classified into artificial reservoir, natural lake or water course. For both the Anatolian and worldwide studies, the concentration of metals analysed for bioaccumulation and the type of tissue examined were then recorded, even though metals present at levels below detection (with limits determined depending on the individual study) were excluded from the dataset. This was because the aim of the present study was to provide a ‘quantitative’ rather than a ‘methodological’ review, but also for reasons of parsimony in the presentation of results (e.g. Patiño et al. 2013) and considering that ‘blank’ values would have not affected the outcomes of the meta-analysis. Overall, this ensured equality control and quality assurance of the methods utilised in the reviewed studies. Further, despite early attempts, no grouping of metals into ‘classes’ was made because of inconsistencies in their classification (see Nieboer and Richardson 1980; Duffus 2002; Mahboob 2013). In case of studies with replicated spatial (i.e. site level) and/or temporal (i.e. seasonal, monthly) sampling or dealing with C. carpio scale variants (i.e. scale, mirror) or different sexes, the mean concentration value of a metal was taken for the comparative analyses. The same criterion was applied to studies carried out in the same water body by different authors.

Owing to concentration measurements being reported in different forms (i.e. mg/g, mg/kg, μg/g, μg/kg, ppm) and to μg/g being taken as the reference measure in the present study due to its wider employment throughout the reviewed literature, the following relationship was used for conversion:

$$ \mathrm{ppm}=\mathrm{mg}/\mathrm{kg}=\upmu \mathrm{g}/\mathrm{g}=\mathrm{mg}/\mathrm{g}\times 0.001=\upmu \mathrm{g}/\mathrm{g}=\mathrm{ng}/\mathrm{g}\times 1000 $$

Similarly, being wet weight most commonly reported across studies than dry weight, the following relationship was used (after Sorensen 1991):

$$ \mathrm{Wet}\ \mathrm{weight}=\mathrm{Dry}\ \mathrm{weight}\times 0.33 $$

Finally, whenever possible, the maximum allowed limit for concentration of a metal was obtained from the Turkish Food Codex (TFC 2002, 2011) and, if not provided therein, from FAO guidelines (Nauen 1983), with values averaged over all countries for which it was reported. Notably, except for As, Cu and Zn, for which the limits provided in the reviewed studies reporting such values were in agreement with those specified in the TFC, those for other metals were defined based on a combination of the latter whereby the selected value was set following a ‘precautionary approach’.

Statistical analysis

Differences in mean concentration amongst waterbody types (i.e. artificial reservoirs, natural lakes and water courses) and the most widely analysed tissues (i.e. muscle, liver and gill) for the eight most studied metals (i.e. Cd, Cr, Cu, Fe, Mn, Ni, Pb and Zn) on C. carpio in Anatolia were tested by permutational univariate analysis of variance (PERANOVA). This was based on a three-way (fully-crossed) design, which included the fixed factors waterbody type, tissue and metal. Analysis was carried out in PERMANOVA + v1.0.1 for PRIMER v6.1.11 (Anderson et al. 2008) following normalisation of the data, using a Euclidean distance and 9999 permutations of the raw data (Anderson and Robinson 2001), and with statistical effects (including a posteriori pair-wise comparisons, in case of significance) evaluated at α = 0.05. Briefly, the advantage of PERANOVA compared to traditional parametric analysis of variance is that the stringent assumptions of normality and homoscedasticity, which proved very often unrealistic when dealing with ecological datasets, are ‘relaxed’ considerably.

For the worldwide comparison, the mean values of the eight most studied metals (as above) were assessed against those for C. carpio from water bodies worldwide by means of quartile analysis (computed under Excel® 2013 as: Q0 = minimum; Q1 = lower quartile or 25th percentile; Q2 = second quartile or 50th percentile; Q3 = third quartile or 75th percentile; Q4 = maximum). Notably, quartiles were computed based on the mean concentration values of each metal from the water bodies worldwide. The waterbody-specific concentration values for C. carpio in Anatolia were then ranked from 0 to 4 according to the quartile into which they fell.

Results

Based on the 43 studies in total selected for review, bioaccumulation data were retrieved for C. carpio sampled from 42 water bodies across Anatolia (Table 1). In total, 27 metals (Table 1) were analysed for concentration from twelve tissues of carp (Appendix Table A1). Amongst metals, Cd, Cr, Cu, Fe, Mn, Ni, Pb and Zn were the most widely analysed; amongst tissues, the muscle, liver and gill were the most commonly used (Appendix Table A2). Maximum allowed limits were provided by the TFC for As, Cd, Cu, Hg, Pb and Zn, and worldwide by FAO for Cr and Se (Appendix Table A2). Of the eight most studied metals, Cd, Cr and Pb occurred in the muscle, liver and gill in concentrations above or well above the maximum allowed limit in several water bodies, and the same was true for Zn except for the muscle tissue (Fig. 1). On the contrary, Cu was always present in concentrations below limit, whereas no limits were available for Fe, Mn and Ni, whose concentrations varied from (comparatively) high to (comparatively) low across the water bodies in which they were studied. Of the lesser-studied metals for which a limit was available, As occurred above threshold in two of the three water bodies where it was investigated, similar to Se, whereas Hg occurred at concentrations above limit in only one waterbody, and its detection was limited to the muscle tissue (Fig. 2).

Table 1 Metals and metalloids (collectively, ‘metals’) analysed for concentration in various tissues of C. carpio from water bodies of Anatolia (Turkey). For each water body, the corresponding code, type (N = natural lake; R = artificial reservoir; W = water course), latitude and longitude are given
Full size table
Fig. 1
figure 1figure 1

Mean (± SD) concentration of the eight most abundant metals in the muscle, liver and gill tissues of C. carpio from water bodies of Anatolia (codes in Table 1). Whenever provided, the maximum allowed limit for concentration is indicated by a dashed line after Turkish Food Codex (TFC 2002, 2011), if available, or as a mean value after FAO (Nauen 1983) (corresponding metal marked by an asterisk; see also Appendix Table A2)

Full size image
Fig. 2
figure 2

Mean (± SD) concentration of three metals in the muscle, liver and gill tissues of C. carpio from water bodies of Anatolia (codes in Table 1). Limits as in Fig. 1

Full size image

Across Anatolia, amongst the metals for which maximum allowed limits were available, only Cu and Hg always occurred below threshold at overall mean concentrations (i.e. averaged over muscle, liver and gill tissues). Conversely, concentrations of Cd and Pb were above limit in 18 and 20 water bodies, respectively (out of the 42 in total investigated), that of Cr and Zn in six and ten water bodies, respectively, and that of As and Se in one water body (namely, Bafra Balık Lakes) in each case (Fig. 3). Overall, in 27 of the water bodies examined (equal to 67.5 % of the total) C. carpio was found to have one or more metals in overall mean concentrations above limit. Specifically, Bafra Balɪk Lakes and Işıklı Reservoir were the water bodies with the largest number (n = 4) of such metals at critical concentrations, namely As, Cd, Se and Zn for the former and Cd, Cr, Pb and Zn and for the latter. Of the other water bodies, six (22 % of total) had three metals, ten (37 %) two metals and the other nine (33 %) one metal above limit (Fig. 4).

Fig. 3
figure 3

Water bodies of Anatolia with above-limit mean concentrations of metals in tissues (muscle, liver and gill combined) of C. carpio. Limits as in Fig. 1

Full size image
Fig. 4
figure 4

Water bodies of Anatolia with one to four metals in C. carpio tissues (muscle, liver and gill combined) above maximum allowed limit for concentration

Full size image

For the eight most investigated metals, there were no significant differences in their mean concentration amongst waterbody types, but there were differences amongst tissues depending on the metal (Table 2). Specifically, the mean concentrations of Cu, Fe, Pb and Zn (with the latter two metals above limit) were always significantly higher in the liver relative to muscle tissue and, except for Cu, the same was true for the gill relative to muscle tissue (Fig. 5). Conversely, there were no significant differences in the concentration amongst tissues for Cd, Cr, Mn and Ni.

Table 2 PERANOVA results showing differences in the concentration of the eight most abundant metals in the muscle, liver and gill tissues of C. carpio from water bodies of Anatolia. Statistically significant effects (α = 0.05) are highlighted in italic, including those for a posteriori pair-wise comparisons (# = permutational value based on 9999 permutations). See also Fig. 4
Full size table
Fig. 5
figure 5

Differences in mean (± SD) concentration for four of the eight most studied metals in the muscle, liver and gill tissues of C. carpio from water bodies of Anatolia. Limits as in Fig. 1. Black bars: mean concentration above limit; Pair-wise statistically significant differences are indicated (see Table 2)

Full size image

Worldwide, based on 56 studies in total (Table 3), bioaccumulation data for C. carpio were retrieved for 102 water bodies across 23 countries. Upon comparison with the worldwide dataset (Table 3): for Cd, eight out of 18 Anatolian water bodies had concentration values (muscle, liver and/or gill tissues combined) above Q2, of which three within Q3 and the other five within Q4; for Cr, five out of six Anatolian waterbodies had concentration values above Q2, of which one within Q3 and the other four within Q4; for Pb, nine out of 20 Anatolian water bodies had concentration values above Q2, of which four within Q3 and the other five within Q4; for Se, only one water body (namely, Bafra Balɪk Lakes) had a critical value within Q0; and for Zn, four out of ten Anatolian water bodies had concentration values above Q2, of which three within Q3 and the other one within Q4 (Fig. 6).

Table 3 Comparative table for the concentration (μg/g) of metals in carp from water bodies of Anatolia (Turkey) relative to waterbodies worldwide. For each of the most widely studied metals, the mean concentration is provided only if above the maximum allowed limit, otherwise indicated by a ‘<’ value (see Appendix Table A2). For water bodies of Anatolia, the quartile range (0 = minimum; 1 = lower quartile or 25th percentile; 2 = second quartile or 50th percentile; 3 = third quartile or 75th percentile; 4 = maximum) is given based on the mean concentration values for the water bodies worldwide
Full size table
Fig. 6
figure 6

Water bodies of Anatolia with concentration of metals in C. carpio tissues (muscle, liver and gill combined) above maximum allowed limit relative to critical values worldwide. Quartiles as defined in Table 3

Full size image

In total, in 16 out of the 42 water bodies examined (equal to 38 %) C. carpio was found to have one or more metals in a concentration (muscle, liver and/or gill tissues combined) within or above Q3 (Fig. 7). Specifically, Beyşehir Lake together with Demirköprü and Işıklı reservoirs were the water bodies with the largest number (n = 3) of metals at critical concentrations, and specifically Cd, Cr and Pb. Of the other water bodies, six (31 % of total) had two metals and the remaining eight (50 %) one metal within or above Q3.

Fig. 7
figure 7

Water bodies of Anatolia with one to four metals above maximum allowed limit for concentration in C. carpio tissues (muscle, liver and gill combined) relative to critical values worldwide. Quartiles as defined in Table 3

Full size image

Discussion

Similar to other biological aspects of C. carpio investigated for Turkey (Vilizzi et al. 2013, 2014a, b, 2015a, b), the dense matrix of studies available at the country level has provided an extensive dataset from which to draw general conclusions about the extent of bioaccumulation of metals in C. carpio tissues. Cd, Cr and Pb were the toxic metals that occurred in several cases at concentrations not only above maximum allowed limits but also comparatively higher relative to other water bodies worldwide. Other toxic metals such as As, Hg and, possibly, Ni also were present in some cases at potentially hazardous concentrations. Whereas, the essential metals Cu, Fe, Mn, Se and Zn were detected at various concentrations, with the latter two in some cases above limit. Overall, the present findings warrant careful evaluation given that C. carpio is caught throughout the country (Harlioğlu 2011) and represents in most areas an important food resource (Turkish Statistical Institute 2014).

Amongst the metals regarded as toxic (and studied more extensively in Turkey): cadmium (Cd) injures the kidney and causes symptoms of chronic toxicity, including impaired kidney function, poor reproductive capacity, hypertension, tumours and hepatic dysfunction (Waalkes 2000); chromium (Cr + 6), despite being in some cases a nutrient essential to glucose metabolism, at higher doses may cause severe environmental and public health problems and is considered a carcinogenic and mutagenic substance for humans (Rowbotham et al. 2000); lead (Pb) is known to be toxic for the brain, kidney and reproductive system and can also cause impairment in intellectual functioning, fertility, miscarriage and hypertension (Flora 2002; Yaman 2006); arsenic (As) can cause acute and chronic toxicity and turn minor disorders into cancer and even death (Bhattacharya et al. 2007); mercury (Hg) is highly toxic in metabolically active tissues causing high ecotoxicological risks (Pelletier 1995); finally, nickel (Ni), despite its being a moderately toxic element, when present in high concentrations in food may be responsible for a number of diseases (Schiavino et al. 2006).

Amongst the metals regarded as essential (and studied more extensively in Turkey): copper (Cu) is known to play a role in the activation of some important enzymes, so that changes in concentration may cause various metabolic alterations (Ozen et al. 2002); iron (Fe) plays an important role in fighting anaemia, which is not uncommon in children and women across Turkey (Walter et al. 1998); manganese (Mn) is a vital trace metal for the functioning of many organic systems where it plays a major role in the regulation of blood sugar and reproduction, digestion, and bone growth and also acts as a cellular antioxidant (Aschner and Aschner 2005); selenium (Se) has attracted attention because of its antioxidant properties, thereby protecting cells from damage (Sorensen 1991); finally, zinc (Zn) is an important trace element in human metabolism and nutrition and plays a major role in the functioning of many biochemical processes (Scherz and Kirchhoff 2006).

Contrary to ponds and water courses, which are generally subjected only to limited angling activities and exploitation mainly related to sport fishing, reservoirs and large natural lakes in Turkey are usually targeted by professional fishing cooperatives (but also amateur fishermen), with the rate of exploitation becoming more intense in the vicinity of highly populated areas. This is the case of e.g. Almus Reservoir (Zengin and Buhan 2007), Keban Reservoir (Celayir et al. 2006; Orsay and Duman 2005, 2008b; Yüngül et al. 2012), Menzelet Reservoir (Alp et al. 2003), Seyhan Reservoir (Ozyurt et al. 2004; Çiçek et al. 2006; Mete and Yüksel 2014) and Lake Eğirdir (Balık et al. 2007). All these water bodies support productive C. carpio fisheries and were flagged in the present study as having critical levels of certain toxic metals in C. carpio tissues.

Based on the current review, an even more critical situation was identified for Bafra Balɪk Lakes and Işıklı Reservoir (relative to Turkish water bodies) and for Beyşehir Lake (relative to water bodies worldwide), and this was due to elevated concentrations of both toxic and essential metals. Notably, Işıklı Reservoir is located in populated Denizli Province and supports cooperative-based fishing exploitation of C. carpio (Eren and Tenekecioğlu 2006), whereas Beyşehir lake (which is located in the city of Konya) is known to support very large catches of C. carpio (Balık and Çubuk 2004).

Bafra Balık Lakes also were flagged in the present study as yielding high concentrations of toxic As (but also essential Se and Zn) in C. carpio tissues. This lake system consists of 33 small water bodies located in Samsun Province, where C. carpio also represents a favourite local dish. Contrary to artificial reservoirs, with their typical V-shaped profile and high depth resulting in a restricted littoral zone area, this type of water bodies provides for extensive shallow and vegetated areas ideal for successful spawning of C. carpio (Balon 1995). In this respect, it has been argued that, if successful stocking of C. carpio and related productive fisheries are to be achieved, shallow (natural) lakes in general should be targeted as potential alternative waterbodies to reservoirs (Gaygusuz et al. 2015). As Bafra Balık Lakes does represent one such potential area in a country with overall limited availability of shallow water bodies, the finding of elevated metal concentrations in the feral population of C. carpio in these lakes raises concerns not only about the feasibility of suggested potential stocking alternatives but also about the safety of C. carpio consumed locally as food.

The significantly higher concentrations of some metals (namely, Cu, Fe, Pb and Zn) in the liver (but also gill) relative to the muscle tissue are in line with findings from other studies, which indicate that the liver and gill are metabolically active and accumulate metals in higher levels relative to the muscle, not only in C. carpio (e.g. Velcheva 2006; Alhashemi et al. 2012; Ardakani and Jafari 2014) but more generally in fish (Allen 1994, 1995; Spehar et al. 1998). This is because, unlike other tissues, the muscle does not come into direct contact with metals as it is covered externally by the skin and also does not function as an active site for detoxification (hence transport of trace metals to other tissues). Further, as regards food processing, this finding is noteworthy as it may have important implications for health and food safety should the preparation of meat balls and sausages from C. carpio involve the utilisation of body parts other than the muscle (fillet).

Unlike the many studies (both Turkey-based and worldwide, and reviewed herein) on C. carpio bioaccumulation that have focused on individual water bodies, to the author’s knowledge, the present review has been the first one to provide a country-level synthesis and evaluation of the topic. This is important if guidelines are to be provided for the management of sustainable and healthy fisheries and related legislation is to be amended/implemented. To this end, a similar meta-analytical approach as the one used in the current study is suggested for other countries where C. carpio represents an important resource for fisheries. This could be the case for e.g. Iran, where in the last two decades the per capita consumption of fish (including C. carpio) has increased sharply (Ebrahimi and Taherianfard 2011), and for which the second largest number of bioaccumulation studies (after Turkey) was retrieved in this review (Table 3).

At the country level, the present findings indicate that there is a pressing need to re-evaluate the status of C. carpio fisheries in most water bodies, and especially those located in the proximity of highly populated areas due to potentially excessive levels of bioaccumulated metals that may ultimately pose a risk to human health. As the current TFC does not include limits for toxic Cr, the value of 1.0 μg/g suggested in the present study may be included in future legislation and the same could apply to the value of 1.15 μg/g for (essential) Se. Finally, within a broader international context, the current synthesis and evaluation study on bioaccumulation of metals in (edible) C. carpio in Turkey will help contribute to the improvement of the country’s food safety policy, so as to align with e.g. expected European Commission standards (European Commission 2013).