Portable Smoking Ovens: What Are the PAH Levels in Grilled and Smoked Rainbow Trout?

Abstract

Fish smoking is one of the oldest preservation methods and has traditional forms in most parts of the world, each with distinct variables. The presence of PAHs in smoked fish is of great concern for producers and consumers alike, as PAHs have great negative effects on human health. This study aimed to determine the physicochemical characteristics and polycyclic aromatic hydrocarbons (PAHs) content level in smoked rainbow trout Oncorhynchus mykiss (Walbaum, 1792) processed using a commercial portable grilling and smoking oven. Sawdust of hardwood (beech, cherry, sour cherry, walnut, and plum) and softwood (fir, willow) were used to produce aromatic essences. The principal component analysis (PCA) revealed that different tree species tend to accumulate different PAHs in different concentrations. In the case of the generalized additive model (GAM) analyzing fish meat, fir tree reduced the concentration of PAHs in fish meat. When GAM analyzed fish skin, cherry, sour cherry, and willow trees significantly reduced the concentrations of PAHs in fish skin compared to beech trees. Furthermore, the results regarding the skin of the fish suggest that it acts as a protective barrier, trapping smoke particulates and reducing the penetration of PAHs into the meat. The present method clearly shows that, at least in the case of Benzo[a]pyrene, it is safer than traditional methods of smoking. This highlights the need for further research into the physicochemical properties of fish tissue and their impact on PAH accumulation.

1. Introduction

The consumption of fish by humans’ date back to the upper paleolithic period, 40,000 to 50,000 years ago [1], with other sources mentioning an earlier period of almost 2 million years ago [2]. Fish is a source of high-quality protein, omega-3 fatty acids, minerals, and other nutritional compounds, which provide several human health benefits [3]. Thus, it is understandable why fish is one of the most consumed foods worldwide, with approximately 20 kg per individual in 2016, stabilizing afterward, with some variations [4]. In 2023, the fisheries production of fish was 90.6 million metric tons, while aquaculture production surpassed it, reaching 96 million metric tons, totaling a fish production of over 186 million metric tons [4]. However, fish is one of the most perishable foods, and the extension of its shelf-life is a subject of continuous study [5].

Smoking is one of the oldest preservation methods for meat and fish. Fish smoking has different forms in most parts of the world, each with distinct variables [6,7]. Smoked fish is an established traditional food in Eastern and Northern Europe, and it is gaining increasing popularity in Western Europe. Fish smoking, apart from the advantages it conveys to the final product, such as longer shelf-life, specific and desired organoleptic properties [8,9], longer life of eicosapentaenoic and docosahexaenoic acids [10], does have some downsides, such as the deposition and accumulation of polycyclic aromatic hydrocarbons (PAHs), nitrosamines, and heterocyclic amines in the final product [11]. For smoking, different types of wood may be used. Hardwood produces a milder smoke, with a more subtle flavor, and burns longer. Softwood confers a more intense smoke and flavor but generally produces more harmful compounds. Hardwood is usually the preferred choice for smoking fish [12].

The presence of PAHs in smoked fish is of great concern to producers and consumers alike, as PAHs have great negative effects on human health with known carcinogenic effects. Long-term exposure can lead to respiratory issues, cardiovascular diseases, and neurological damage. They can also affect the immune system and cause genotoxicity [13,14]. PAHs are a result of wood combustion, depositing in the smoked product. Many variables influence the levels of PAHs in smoked products: the species of fish used, the processing of fish until smoking, the wood material used, the level of heat and time of exposure, the packaging after smoking, and many others [15,16,17].

By modern standards, the method of preservation should not only maintain fish quality and freshness but also respect consumer concerns regarding food safety and the sustainability of resources [18,19]. Moreover, in some cases, there is a cultural dimension to the method of preservation, specific to a people or a region [20]. When employing traditional methods of preservation, there needs to be a balance between the cultural dimension and the other aspects related to regulations regarding consumer safety and other consumer concerns. Nowadays, there is an increasing awareness of self-sustainability concerns, with people having more access to information and a desire to grow, capture, prepare, and cook their food. For example, the concept of “farm-to-fork” can have a shorter chain: catch a fish and benefit from its great nutritional properties by using a portable grilling and smoking oven, thus providing a possibly healthier and similar experience to traditionally smoked fish.

This study aims to determine the physicochemical characteristics and PAHs levels in smoked rainbow trout Oncorhynchus mykiss (Walbaum, 1792) processed using a commercial portable grilling and smoking oven, a method not yet explored in the literature. Specifically, we seek to answer the following questions: Will wood types (softwood and hardwood) and wood species (fir, willow, beech, cherry, sour cherry, walnut, and plum) used for smoking fish influence the concentrations of PAHs in fish meat and skin? Will different fish tissues (meat and skin) accumulate PAHs differently due to direct contact with smoke? Is this method of grilling and smoking an alternative to traditional smoking methods from organoleptic and health perspectives?

2. Materials and Methods

2.1. Fish and Experimental Protocol

The fish specimens used in this study were sampled from a trout farm near Cluj-Napoca City, Cluj County, Romania, and were represented by twenty-one specimens of rainbow trout (mean weight of the fish was 423.5 ± 16.48 g). We followed the technological flow before smoking described in a previous paper by Sava et al. [16] with modifications, which consisted of the following stages: harvesting, immediate mechanical stunning and bleeding, evisceration, washing (to remove mucositis and surface impurities adhering to the fish surface, as well as eliminate viscera), salting (dry salting for 16 h under refrigeration conditions at 2–4 °C), a second stage of washing (to remove excess salt), drying (the washed fish is left to dry for up to 4 h at 8–9 °C). The Mohr method was used to determine the salt content of the smoked trout [21]. The amount of salt remaining in the products was a maximum of 3%.

For the smoking stage, a portable grilling and smoking oven EnergoTeam (Figure 1) 40 × 24 × 10 cm (Energofish Ltd., Budapest, Hungary) was used. The oven is made of stainless steel. It comprises from bottom to top: two separate spirit lamps for heating; a main compartment in which sawdust or other smoking material is placed at the bottom; a stainless-steel plate to partially cover and guide the smoke; a grill for the fish; a lid with a handle that can regulate airflow. This oven has the following advantages: it is user-friendly, easy to transport and clean, time-efficient and low-cost. The main disadvantage is its low production capacity, making it unsuitable for industrial purposes and only for personal use.

The oven was preheated for ten minutes until the inside temperature reached 120 °C. Different types of fuel can be used for heating. In the present study, the fuel consisted of ethanol. Three fish were placed on the grill equally spaced for each sawdust test. Sawdust with an average humidity of 10–12% from hardwoods such as beech, cherry, sour cherry, walnut, and plum, as well as softwoods including fir and willow, were used to produce aromatic essences. For this type of method using grilling and smoking ovens, 5 g of sawdust is sufficient for producing the specific aroma and taste. Hot smoking at 120–140 °C for up to 30 min was maintained until the internal temperature of the fish reached 90 °C. The temperature of the oven was monitored with a DVFS-T500 thermometer (Dvorsons FSE, Sausalito, CA, USA). The internal fish temperature was monitored by a traceable alarm thermometer/timer (Fisher Scientific, Pittsburgh, PA, USA). The same smoking process was applied to all wood species used to eliminate any potential variations in the results that could arise from differences in the smoking procedure. Between each smoking batch, the smoking chamber was cleaned, washed, and dried. There were three fish per batch, totaling seven batches, with one for each type of wood. After grilling and smoking, packaging and storing were performed under refrigeration conditions at 2–4 °C until further analyses. The smoked meat and skin from all specimens from each smoking batch were manually separated. The smoked meat and skin samples were each homogenized per batch, macerated using a mortar and pestle, and prepared for further analyses.

2.2. Physicochemical and Polycyclic Aromatic Hydrocarbons (PAHs) Analysis

The protein, total lipids, and moisture content were conducted using FT-NIR Tango from Bruker (Billerica, MA, USA) [22]. Samples were measured directly without extraction. The method parameters included a measurement time of 64 s, a resolution of 16 cm⁻¹, and a rotating scan type. The results were compared to calibration curves provided by Bruker, with a bias correction applied to the calibration curve. This correction was calculated by comparing the results obtained from the FT-NIR with those from the standard methods. The final results were expressed as a percentage.

The HPLC method was used to determine the content of PAHs in the traditionally smoked trout meat, according to ISO Method SR EN ISO 17993:2006 [23]. A Perkin–Elmer 200 Series high-performance liquid chromatograph (HPLC) with an FLD detector (PerkinElmer Inc., Waltham, MA, USA), with an Inertsil ODS-P 5 µm, 4.6 × 150 mm, kept at 24 °C, was used. This setup facilitated the separation of 15 polycyclic aromatic hydrocarbons (PAHs), including naphthalene, acenaphthene, fluorene, phenanthrene, anthracene, fluoranthene, pyrene, benzo(a)anthracene, chrysene, benzo(b)fluoranthene, benzo(k)fluoranthene, benzo(a)pyrene, dibenzo(a,h)anthracene, benzo(g,h,i)perylene, and indeno(1,2,3-cd)pyrene. The mobile phase consisted of a gradient of water and acetonitrile and a time-programmed FLD detector used for the detection of the 15 PAHs. Ten grams of homogenized fish samples were saponified using a 50 mL KOH 0.4 M solution in ethanol and water (9:1) in an ultrasonic bath at 60 °C for 30 min. After filtration (on an ash-free cellulose filter), the samples were extracted twice using 15 mL hexane. The collected supernatant was then passed through a 1 g Florisil column (the column was washed with 15 mL hexane) and then concentrated in a stream of nitrogen. The sample was then redissolved in 1 mL of acetonitrile and injected into the HPLC system. All used solvents were HPLC grade from Merck (Merck KGaA, Darmstadt, Germany), the KOH (pellets), and Florisil (0.125–0.250 mm) from Supelco (Saint Louis, MO, USA). The content of PAHs in the samples was quantified and expressed in nanograms per gram (ng/g) [23]. The method was validated for quantitatively determining 15 PAHs, with linear calibration ranges between 0.0005 and 0.050 μg/kg and correlation coefficients (R²) ≥ 0.99. The LOD and LOQ ranged from 0.0001 to 0.00015 and 0.00033 to 0.00045 μg/kg, respectively, based on a signal-to-noise ratio of 3:1 (LOD) and 10:1 (LOQ). Method accuracy (recovery) was evaluated at three spiking levels (low, medium, and high) and ranged from 75.22 to 92.03%. No significant matrix interferences were observed, confirming the specificity of the method.

2.3. Statistical Analysis

Considering our hypotheses, we formulated several research questions: 1. Do different wood species and types have different concentrations of PAHs? 2. Do the concentrations of different PAHs in wood species used for smoking influence the concentrations of PAHs in fish meat and skin? 3. Are there correlations between the smoked fish’s total lipids, protein, and moisture content and different PAHs? 4. Which PAHs are most likely to accumulate in fish meat and skin, and can we avoid these by using some species or types of wood for smoking?

The chemical analyses produced a dataset of concentration values of different PAHs in wood (wood_PAH) according to wood species (wood_sp) and wood type (softwood or hardwood), the concentration of PAHs in fish meat (meat_PAH), and in fish skin (skin_PAH). The variables were tested for normal distribution with the Shapiro–Wilk test [22]. Since neither of the variables was found to be normally distributed, we transformed the data with the formula: transformed variable = ln (variable + 1) + 1. The resulting transformed values fitted a gamma distribution. The threshold for significance was considered to be at p = 0.05 in all analyses.

First, we performed an exploratory PCA (RStudio, MASS, and ggbiplot packages) to see whether different wood types and wood species are associated with different concentrations of PAHs in wood. Subsequently, we performed a Spearman rank correlation analysis between total lipids, protein, and moisture percentage in the whole fish and different concentrations of PAHs in meat and skin, respectively (cor.test function, corrr package, RStudio).

Secondly, we fitted a generalized additive model (GAM) to explore the relationships between the concentration of PAHs in wood vs. fish meat and wood vs. fish skin. The first attempt was to elaborate a model with the two variables (wood_PAH as the independent variable with meat_PAH and skin_PAH as the dependent variables, respectively). Since the resulting models explained just a small fraction of the variation in the dependent variables (27–33% in fish meat and 17–19% in fish skin), we added the parameters: wood_sp, PAH type, and the interaction between wood_sp and wood_PAH to the GAM. The analysis was performed in RStudio (2024) with the GAM—function (mgcv package). The models were tested (gam.check function in RStudio) with the generalized cross-validation method (GCV); the optimization method was outer Newton; distribution was set to Gamma [24].

3. Results

The proximate composition of the rainbow trout processed in this study is presented in

Table 1. Total lipids, moisture, and protein content in grilled and smoked rainbow trout.

Chemical Analysis (Whole Fish) %	Beech	Cherry	Sour Cherry	Plum	Walnut	Fir	Willow
Total lipids	8.4 ± 0.9	11.7 ± 2.5	11.7 ± 1.3	7.6 ± 2.8	9.1 ± 1.6	9.7 ± 1.9	11.77 ± 2
Moisture	64.5 ± 8.4	65.7 ± 6.1	65.2 ± 10.6	61.5 ± 7.9	63.1 ± 9.7	63.1 ± 7.6	64.5 ± 8.7
Protein	20.1 ± 2.3	19.4 ± 2.7	20.5 ± 2.2	21.9 ± 1.9	21.7 ± 3.1	21.0 ± 2.1	21.6 ± 2.4

. The percentages of total lipids, moisture, and protein content in the fish meat from the specimens used in the smoking experiment show similar levels for these main parameters, with no significant differences observed depending on the type of wood used, indicating consistent composition across the samples after processing.

The types of PAHs identified through chemical analyses are summarized in Supplementary Table S1. Benzo(a)pyrene was the only PAH that was not found in meat at all. All other identified PAHs were found in meat at various concentrations. We found all identified PAHs in skin tissue at various concentrations.

Wood type (hardwood or softwood) was not associated with different patterns of PAH concentrations. The PCA analysis revealed that different tree species tended to accumulate different PAHs in different concentrations (Figure 2).

The PC axes 1 and 2 explained a total of 90% variation in the data. On PC1, the effects caused by the concentrations of Anthracene (A2) were most different from those of Chrysene (C1), Bezo(g,h,i) perylene (B5), and Indeno(1,2,3-cd)pyrene (I1). On PC2 the effects caused by the concentrations of Phenanthrene (P1), Anthracene (A2), Fluoranthene (F2), and Pyrene (P2) were most different from those of Benzo(b)fluoranthene (B2) and Benzo(a)pyrene (B4). Willow, in particular, differed from the other wood species and accumulated the highest concentrations of Anthracene (A2), Phenanthrene (P1), Pyrene (P2), and Fluoranthene (F2) (Figure 2).

Beech, plum, cherry, fir, and walnut trees tended to accumulate similarly different PAHs. Sour cherry, on the other hand, had higher concentrations of Naphthalene (N1), Benzo(a)pyrene (B4), Benzo(b)fluoranthene (B2), Benzo(k) fluoranthene (B3), and Fluorene (F1) (Figure 2).

We found significantly strong negative correlations between total lipids % and skin concentrations of Bezo(b)fluoranthene (Rho = −0.875, p = 0.033), Benzo(k)fluoranthene (Rho = −0.935, p = 0.007), Benzo(g,h,i)perylene (Rho = −0.860, p = 0.030), and Indeno(1,2,3-cd)pyrene (Rho = −0.875, p = 0.033); and moisture % and skin concentrations of Chrysene (Rho = −0.909, p = 0.008) and Benzo(g,h,i)perylene (Rho = −0.849, p = 0.026).

The result of the first GAM indicated that wood concentration of PAHs has a significant effect on the concentrations of PAHs in fish meat (Estimate = 0.249, t value = 9.354, p < 0.001; smooth term: EDF = 2.161, F = 14.5, p < 0.001; adjusted R² = 0.271, deviance explained = 32.9%) and skin (Estimate = 0.587, t value = 13.28, p < 0.001; smooth term: EDF = 1.358, F = 11.21, p < 0.001; adjusted R² = 0.176, deviance explained = 19.4%). However, only a small percentage of the variation in the data (27–33% in the case of fish meat and 17–19% in the case of fish skin) was explained by the model. After adding to the GAM the parameters wood_sp and PAH type, the model explained a much higher part of the data variation: 87–91% for fish meat, and 81–86% for fish skin.

In the case of the GAM analyzing fish meat, fir tree reduced the concentration of PAHs in fish meat compared to beech (reference term); Benzo(a)pyrene and Dibenz(a,h)anthracene reduced the concentration of PAHs in fish meat; the smooth term was significant (F = 5.017, p = 0.028) in the case of fir wood PAH concentration vs meat PAH concentration, and indicated a near-linear relationship between wood concentration and meat concentration of PAHs (EDF = 1, Figure 3a). The relationship is negative, indicating that a higher concentration of PAHs in fir wood predicts a lower concentration in fish meat. The plot in Figure 3a indicates that this relationship is significant and valid for lower-value intervals of the wood PAH concentrations (rug plot). The model check indicated that all relationships were linear or near-linear; the model fitting was successful and the solutions stable (Supplementary Table S2).

In the case of the GAM analyzing fish skin, cherry, sour cherry, and willow significantly reduced the concentrations of PAHs in fish skin compared to beech (reference term); walnut significantly increased the concentration of PAHs in fish skin compared to beech; Benzo(a)pyrene reduced the concentration of PAHs in fish skin, while Chrysene, Fluoranthene, Fluorene, Naphthalene, and Phenanthrene significantly increased the concentrations of PAHs in fish skin; the smooth terms were significant in the cases of beech (EDF = 1, F = 4.094, p = 0.047), plum (EDF = 3.505, F = 7.496, p < 0.001), and walnut (EDF = 1.663, F = 3.935, p = 0.025) and indicated near-linear and non-linear relationships. However, the relationship between wood_PAH for beech with skin_PAH has a very small slope and is negligible (Figure 3b). On the other hand, the non-linear relationships between wood_PAH for plum and walnut with skin_PAH showed very little effect at lower values of wood_PAH, and a stronger but more uncertain effect at higher values (Figure 4a,b). The model check indicated that the model fitting was successful and the solutions stable (Supplementary Table S2).

4. Discussion

PAH can be formed in food during processing and home preparation methods like smoking, drying, roasting, baking, frying, or grilling. Several factors, including smoking temperature, wood type, and potential contamination of the wood source influence the formation of polycyclic aromatic hydrocarbons (PAHs) in smoked fish [25,26,27,28,29,30]. PAHs are predominantly formed during pyrolysis and incomplete combustion at temperatures between 200 and 500 °C [31,32], with lower temperatures, such as those used in this study, limiting the formation of heavier PAHs like benzo[a]pyrene [11]. The European Commission set maximum levels of 5 and 30 μg/kg for benzo(a)pyrene (BaP) and the sum of PAH4 (sum of Benzo[a]pyrene, Benzo[a]anthracene, Benzo[b]fluoranthene, and Chrysene), respectively, in fish meat and meat products that have undergone a heat treatment potentially resulting in formation of PAH, i.e., smoking, grilling, and barbecuing [33].

Among the numerous PAHs, benzo[a]pyrene is the most extensively studied and is commonly used as a marker for PAHs in both ambient air and food [34]. In our study, benzo[a]pyrene was the only PAH not detected in meat samples. These compounds’ absence or reduced levels in fish meat can be attributed to insufficient thermal energy for their synthesis or deposition. According to the European Scientific Committee on Food’s assessment of the risks associated with polycyclic aromatic hydrocarbons (PAHs) in food [34], benzo[a]pyrene levels as high as 200 µg/kg have been reported in smoked fish and meat. In grilled or barbecued fish meat, concentrations of up to 130 µg/kg have been documented, whereas the typical background levels in uncooked foods generally range from 0.01 to 1 µg/kg. In our study, the highest total PAH4 concentration in meat was 0.37 µg/kg, while in skin samples, it reached 15.45 µg/kg for grilled and smoked fish prepared with plum tree sawdust. These values are significantly lower than the limits set by the European Commission [33] and those mentioned by the European Scientific Committee on Food [34] (with lower and upper bounds for the mean concentration ranging from 1.90 to 20.89 µg/kg). The literature reports highly variable quantitative data for PAHs in smoked meat and fish products. These discrepancies can be partially attributed to differences in analytical methods used to assess PAH presence, however, the primary source of variation lies in the differences in smoking procedures.

The type of wood used also plays a critical role [19,35]. Surprisingly, in our study, fir wood, which was associated with lower PAH concentrations in fish meat, may release fewer volatile organic compounds during combustion compared to hardwoods like beech, which produce a steady burn but may generate compounds that favor PAH formation under similar conditions as ours [29].

Furthermore, the results regarding the skin of the fish suggest that it acts as a protective barrier, trapping smoke particulates and reducing the penetration of PAHs into the meat. This highlights the importance of retaining the skin during the smoking process but removing it before human consumption. For example, in cherry tree-smoked fish, our sample showed a low ∑PAH17 of 5.59 μg/kg in the meat, while the skin had a significantly higher ∑PAH17 of 32.79 μg/kg. These findings are consistent with earlier studies that found high PAH concentrations in fish skin, suggesting that retaining the skin during smoking, managing combustion parameters, and setting optimal smoking temperatures play a pivotal role in lowering PAH formation [17,36,37].

A less-explored but equally important factor is the potential contamination of the wood source itself. Wood can accumulate environmental contaminants such as pesticides, heavy metals, and PAHs from industrial emissions, preservatives, and pollutants from the surrounding environment, subsequently influencing the PAH profile in smoked products. For instance, wood sourced from areas with high levels of air or soil pollution may already contain measurable PAHs or other organic pollutants that can volatilize during smoking. This contamination could explain the observed variability in PAH concentrations among different wood species and underscores the need for stringent sourcing and quality control of smoking materials [32,38]. The development of pre-treatment methods for wood, such as washing or thermal treatment, could significantly reduce the presence of pre-existing contaminants.

The interaction between PAHs, as observed in this study, further complicates their behavior during smoking. While lighter PAHs such as fluorene and naphthalene are more volatile and tend to be present in higher concentrations due to their lower thermal stability, heavier PAHs like Benzo[a]pyrene are more thermally stable but may undergo partial decomposition during the process, leading to a negative correlation with overall PAH levels. This suggests that lighter PAHs may act as markers for incomplete combustion, whereas heavier PAHs are more sensitive to variations in temperature and oxygen availability [39,40].

Another critical consideration is the role of lipid content in fish meat and skin in absorbing and retaining PAHs. Fish with higher lipid content may have an increased affinity for lipophilic PAHs, potentially altering the distribution of these compounds between the skin and meat. This is consistent with previous findings, which reported an inverse correlation between lipids and PAH content [29,40]. Triglycerides are likely more effective in the initial stages of PAH absorption, while phospholipids may play a more significant role in the later stages, particularly in cellular uptake and distribution [41]. This highlights the need for further research into the physicochemical properties of fish tissue and their impact on PAH accumulation.

5. Conclusions

The implications of these findings go beyond fish smoking practices, highlighting the need for standardized smoking protocols that minimize health risks linked to PAH exposure. Clinicians and nutritionists recommend that PAH levels should be as low as possible, and ideally absent. The method used in this study is safer than traditional smoking, particularly concerning Benzo[a]pyrene. While grilling and smoking ovens cannot replace traditional smoking for product preservation, they offer a healthier alternative for those who enjoy smoky-flavored fish. Implementing real-time monitoring of smoke composition and temperature could help minimize variability in PAH levels.