The mineral supplements were previously tested for content of dioxins, furans, and dioxin-like polychlorinated biphenyls (PCBs). The study was conducted using fattening pigs that were crosses of multiple breeds (Choice-Genetics breeding line), assigned at random to three groups (control and two experimental groups). The initial body weight of the animals was about 35 kg. Each group comprised 240 individuals, kept in three pens with 80 animals in each pen. The experimental groups were designated A and B, and the control group as C. The animals received a standard complete diet adjusted for age (grower and finisher). The animals from groups A and B received a diet with 1.5% composed smectite sorbents, while group C received a standard complete diet. The animals were fed ad libitum with unlimited access to drinking water. Veterinary care was provided by a veterinarian employed on the farm. The microclimatic conditions on the farm (temperature, relative humidity, and air flow), which were systematically monitored throughout the experiment, were in compliance with animal welfare requirements .
Feed and faeces
Feed and faeces were sampled twice for analysis of chemical composition, i.e. at the start (following a 14-day adaptation period) and end of the experiment. Apparent faecal digestibility was determined by the indicator method using 3 g Cr2O3 per kg of feed as an external marker . Faeces were sampled in the morning from six animals in each group and pooled in a container for three days, after which the samples were dried for determination of the rate of drying, ground, and subjected to chemical analysis. Samples for analysis were weighed out from each portion of thoroughly ground faeces, and the concentrations of nutrients and insoluble ash were determined .
The apparent digestibility (faecal digestibility) of the nutrients was calculated from the following equation: apparent digestibility of nutrients = 100 − (100 × content of indicator in feed/content of indicator in faeces × content of nutrients in faeces/content of nutrients in feed).
Previously prepared feed was sampled from six bags, with randomization of the sample. The diets (n = 3) and faeces samples were analysed using AOAC methods  for the content of dry matter (Method 925.09), crude ash (Method 923.03), crude protein by Kjeldahl’s method (Method 920.87), ether extract by Soxhlet extraction (Method 920.39) and crude fibre (Method 962.09). Total P content was determined according to Polish standard PN-76/R-64781 using a Helios Alpha UV–VIS spectrophotometer (Spectronic Unicam, Leeds, UK). Some of the results obtained for the content of nutrients and minerals in the diets are presented in Ossowski et al. . Faecal digestibility was calculated as a percentage (ATTD), as apparent total tract digestibility of organic matter (OM), crude protein (CP), crude fat (CFat), and crude ash (CA) in the pigs’ diet. All analyses were performed in triplicate. The nutritional value of the pig feed was in compliance with nutritional recommendations .
Analysis of blood parameters
Midway through the experiment, blood samples (10 ml) were collected from pigs from each group with similar body weight (average weight from pen about 70 kg per pig). Blood was sampled in the morning before feeding, from each replicate pen in each group, for analysis of haematological and biochemical parameters. From each replicate group 3 samples were collected, i.e. 9 per experimental group. Blood samples were collected from the cranial vena cava into single-use Vacutest tubes (VacutestKimaSrl.) containing lithium heparin. Within three hours of collection, whole blood was analysed for haematological parameters with a Mindray BC 500 Vet analyser (China). Plasma for analysis of biochemical parameters was obtained by centrifugation of whole blood at 3,000 rpm (603 × g) for 15 min in a laboratory centrifuge (MPW-350R, MPW Medical Instruments, Warsaw, Poland) at 4 °C. Samples were stored at − 80 °C until biochemical analysis. The concentrations of glucose (GLU), total protein (TP), creatinine (CREA), urea (UREA), magnesium (Mg), iron (Fe), and phosphorus (P) in the blood plasma were determined by spectrophotometry using Cormay tests. Activity of alanine aminotransferase (ALT), aspartate aminotransferase (AST), and alkaline phosphatase (ALP) was determined by spectrophotometry using Cormay tests. All spectrophotometric analyses were performed using a Mindray BS-120 spectrophotometer. Class A, G, and M immunoglobulins were determined in the blood plasma using ELISA assays (BiokomELx 808, Poland).
Histological tissue analysis
After the fattening period animals from each group were slaughtered, and samples of the lungs, liver, kidneys, and jejunum were taken. These samples were fixed for 24 h in 10% neutral buffered formalin and absolute alcohol, and then passed through increasing concentrations of alcohol solutions and xylene in a tissue processor (Leica TP-1050) and embedded in paraffin blocks. Histological Sects. 4 µm thick, prepared using a sledge microtome (Leica SR-200) and stained with haematoxylin and eosin, were used for morphological analysis under a light microscope. For the liver samples, histochemical staining was additionally performed for the presence of neutral fats. For these analyses, the sections were fixed in 10% neutral buffered formalin, sliced in a freezing microtome (Cryotome FSE, Thermo Scientific), and stained with Sudan IV according to Daddi . For visualization of glycogen, microscope sections of the liver were fixed in absolute alcohol and stained by the PAS method according to McManus .
Analysis of gas residues
Analysis of toxic volatile substances was based on a method by T. Tankin and J.C. Crockett Butler [‘Blood Alcohol Analysis by Static headspace with Dual FID/Megabore Capillary Columns’, Terry Rankin, Jessie Crockett Butler, Thermo Electron Corporation Application Note: 10,076, Milan, Italy].
Analytical standards: acetone, ethanol, methanol, isopropanol (Multi-Component Alcohol Mix-100, Cerilliant, USA), 1-propanol, 2-butanol, 2-pentanol, bezene, dodecane, ethyl acetate, ethylbenzene, m,p,o-xylene (Supelco, Germany), toluene (Sigma-Aldrich, USA). Deionized water (Millipore, USA).
Sample collection and storage
After slaughter, the following material was collected: blood (1 ml), perirenal fat, liver, kidneys, lungs and brain (all 1 g). Biological samples were stored at -20 °C until analysis (Nowakowicz-Dębek et al. ).
Samples in the amount of 200 μl of blood or 200 mg of tissue and 200 μl of internal standard (0.1 g/l tert-butanol solution in water) were placed in 10 ml headspace glass vials, closely crimped using a cap with a PTFE-silicone seal, and analysed by the HS-GC-FID technique.
Instruments and chromatographic conditions
TRIPlus headspace autosampler (Thermo Electron, United States) coupled to a Trace GC Ultra (Thermo Electron, United States) gas chromatograph equipped with one split/splitless injector (working temp. 200 oC) and two FID detectors (working temp. 250 °C). Chromatographic separation was carried out on a dual column system. Two capillary columns (parallel connection using Y-splitter) were used, i.e. Restek Rtx BAC1 and Rtx BAC2 30 m, 0.53 mm ID, 3 and 2 micron film, respectively (Restek, Bellefonte, PA, USA), working at a gradient temperature: initially 40 °C for 2 min, and then increased (15 °C/min) to 150 °C and held for 2 min. The carrier gas was helium, 18 ml/min. TRIPlus headspace autosampler work conditions: sample incubation 75 °C for 8 min, syringe temp. 80 °C, injection volume 0.3 ml. The operation of the gas chromatograph and the HS autosampler were controlled by X’calibur software, which was also used for data treatment. The analytical procedure was validated using certified analytical standards of volatile hydrocarbons (%RSD, LOD, LOQ, linear fit).
The results of the haematological and biochemical analyses for each group were presented as means M with standard deviation (SD), standard error of the mean (SEM), and p-value. Statistical differences between groups were determined by Tukey’s test for a significance level of < 0.05. The analysis was carried out using Statistica software ver. 12.0 (StatSoft S.A., Tulsa, OK, USA). Means between groups marked with different letters (a, b…) differ significantly at p < 0.05.