A simple liquid–liquid extraction was successfully applied to extract the HZ and CZ and the IS from postmortem samples. No interfering ion current signals were observed at the retention time of this compounds (Fig. 2A). The optimized conditions of the sample preparation that were elaborated in this project provided clear mass spectra with low matrix effect in the tested real forensic samples. In Fig. 2, a sample chromatogram of urine and blood with marked peaks of HZ (eluted at 6.67 min) and CZ (eluted at 6.5 min) is presented (Fig. 2 B and C). The elaborated MRM parameters, fragmentor voltage and collision energies that are suitable for the determination of HZ and CZ (and respective standards) in biological samples were optimized for both proposed methods employing the usage of mass detectors with different ionization sources. The optimal parameters for both instrument types are presented in detail. Table 1 displays the validation parameters of the method employed. All presented values fall within the acceptable range for toxicological analysis of biological materials, in accordance with the recommendations of the German Society of Toxicological and Forensic Chemistry (GTFCh) (Peters et al. 2009). As a result, the method has been successfully implemented for routine toxicological analysis in our laboratory.

The targeted LC–MS/MS chromatograms obtained for blank matrix (A) with isolation m/z precursor ion for HZ (A1) and CZ (A2), scan of an authentic blood sample (B), scan of an authentic urine sample (C), product ion spectra of HZ in authentic human blood; collision energy: a –10, b –20, and c –35 V (D, D1–D3), product ion spectra of CZ in authentic human blood; collision energy: a –10, b –20, and c –35 V (E, E1–E3)

Method 1 was performed using an instrument composed of an HPLC chromatograph coupled with an APCI-QqQ-MS/MS.

The operating platform was checked for its selectivity, linearity, LOQ, precision, accuracy, recovery, and stability. As a result of the selectivity assay, no interfering endogenous substances were observed at the retention times of the analytes and ISTD (see Fig. 3). Concerning the linearity of the results, the quadratic and linear regressions of the peak area ratios versus concentrations were fitted over the concentration range 5–1000 ng/ml for CZ and HZ in human blood. The following calibration curve equations were obtained: y = 2.298011 x—0.023062 for HZ and y = − 0.035042 x2 + 0.843162 x—0.028029 for CZ and the coefficient of determination R2 was calculated as R2 ˃ 0.9998 for both compounds, respectively, showing high repeatability of the composed method and precision of the taken measurements. It is important to underline that the calibration equation calculated for CZ is not linear; however, the results from the analyses of the tested samples were collected from the area of the lowest steepness of the curve. Concerning the LLOQ for both drugs in human blood samples, it was calculated as 0.3696 ng/ mL for CZ and 0.345 ng/ mL for HZ. The LOD for these drugs was equal to 0.1232 and 0.1150 ng/mL, respectively. Precision, accuracy, recovery, and stability results of this method are shown in Table 2. As presented below, the recovery rate of HZ and CZ exceeded 90%, which makes this method a reliable one.

Dynamic MRM chromatograms (Method 1) of blood sample spiked (at concentration 100 ng/ml) with HZ (A1-2), HZ-d8 (B), CZ-d8 (C), and CZ (D1-2)

As it is shown in Table 1 both drugs delivered similar fragments from their molecular ions, namely 166.1 and 201.1, which shows their similar chemical structure and the scaffold they are originating from. The results prove that even if the compounds are of similar chemical nature (similar retention time in the applied chromatographic column), HZ is harder ionizable from CZ as the optimal fragmentation energy was set as 124 V in comparison to CZ, which was easily fragmented and its fragmentation value was determined as 60 V in the used instrument equipped in the APCI source.

Among the elaborated methodological conditions, the optimal collision energy values that were used to observe a clear fragmentation of the molecules of interest were ranging between 12 and 68 V. The fragmentation of CZ to the 166.1 m/z ion was easier and demanded a significantly lower collision energy, namely 12 V in comparison with the 201.1 ion that was obtained at 68 V. The results for CZ were opposite from those recorded for HZ, where the fragmentation of the molecular ion to the m/z of 201.1 was easier and appeared at its highest efficiency at 20 V, in comparison to the voltage of 48 V that was necessary to receive the 166.1 signal. In both cases, the deuteration of a molecule leads to the necessity for the elevation of the fragmentation energy value to obtain similar fragments as in the un-deuterated molecule.

Different analytical parameters needed to be set when using another ionization source. The optimized conditions suitable for the determination of the two drugs, HZ and CZ using the ESI ionization source, were elaborated in a similar manner as for the Method 1. First, the selectivity studies performed on the tested solutions did not reveal the presence of any other interfering endogenous substances that could be washed out from the chromatographic column at a similar time as the analytes and ISTD (see Fig. 4). In the following studies, the linearity of the tested concentrations was studied. As a result, the linear regressions of the peak area ratios versus concentrations were fitted over the concentration range 5–1000 ng/ml for CZ and 5–500 ng/ml for HZ in human blood samples. The obtained calibration curve equations were: y = 0.589749x + 0 for HZ and y = 1.78190x + 0 for CZ. As a result, the calculated R2 coefficients exceeded 0.999 for both compounds assuring a high precision and low standard deviation values of the obtained data. The calculated LLOQ for CZ and HZ in human blood was 1 ng/mL. This result confirms high sensitivity of the elaborated protocol which is particularly necessary when analyzing biological samples. The LOQ for the ESI source is the same as for the APCI ionization method. The detailed parameters of the optimization, like precision, accuracy, recovery, and stability results of this method, are presented in Table 3. As previously, the applied conditions show high recovery rates of the drugs of interest.

MRM chromatograms (Method 2) of blood sample spiked (at concentration 100 ng/ml) with HZ, HZ-d8, CZ-d8, and CZ

The analysis of the fragmentation pattern of HZ and CZ brought similar results as in the APCI-based Method 1. The compounds delivered two distinct fragment ions with m/z of 166.15 and 201.15. The MRM transitions of parent ions to product ions were measured at different energy settings. As a result, the collision energies of − 21 and − 40 were selected as the most favorable for the formation of the 201.1 and 166.15 fragments from CZ, respectively, whereas the energies of − 20 and − 40 were selected for the daughter ions of HZ, respectively. Close values obtained for both drugs confirm their similar chemical character and an alike behavior in the electrospray (see Table 4).

The developed LC–MS method was optimized and validated for the simultaneous determination of HZ and CZ in whole blood samples. A significant number of HZ- and CZ-related intoxications concerning accidental poisonings of infants (Simons 2004), as well as several fatalities due to HZ overdose, either due to accident (McIntyre et al. 2013) or committing suicide (Ma et al. 2007), have been described in the scientific literature. As a result, the determination of the above antihistamines in postmortem samples during the investigation of relative forensic cases is needed. Two methods that are presented in this paper can be used for the simultaneous identification of HZ and CZ in whole blood samples and could be also applied, after proper validation, to plasma, serum, urine or gastric content samples for forensic purposes during investigation of respective cases (Nojavan et al. 2012).

To the best of the authors’ knowledge in the scientific literature, there are only two works that relate to the determination of CZ and HZ in human blood (Gergov et al 2001; McIntyre et al. 2013). However, for those studies five times more blood was required in comparison with the method presented in our paper. In the work published by Gergov et.al. (Gergov et al 2001), a LLOQ (1 ng / ml for HZ and CZ) was achieved but with using a time-consuming two-step liquid–liquid extraction. Another drawback of this work is that five times larger volume of biological material was required for conducting an analysis than in our described methods.

An additional advantage of the method described in this paper is the use of two MRM transitions to identify CZ and HZ. So far, all authors have used only a single transition (de Jager et al. 2002; Eriksen et al. 2002; Gergov et al 2001; Kang et al. 2010; Ma et al. 2007; Song et al. 2005). In forensic toxicology, the internal standard technique is the most common one to determine the quantity of a compound of interest in the sample. The use of deuterated internal standards provides high recovery of marked analytes, comparable to the results of the authors who have applied levoceterizine-D8 (Kang et al. 2010), indometacin (Kowalski and Plenis 2007), phenylalanine (Ma et al. 2007), nebivolol (Dharuman et al. 2011), oxybutynin (de Jager et al. 2002) or HZ (Eriksen et al. 2002). The application of HZ as an internal standard may not be desired for the sample itself as the individuals from the tested groups may also take HZ on a daily basis (see Table 5).

The determination of very low levels of drugs in the biological material is a challenge for forensic and clinical toxicologists. While elaborating new analytical methods, next to achieving the lowest possible LLOQ value, it is important to make it simple and fast. Table 5 sets together analytical methods that were elaborated for the determination of HZ and CZ content in biological samples. As presented in Table 5, the majority of scientific papers rely were on liquid chromatography coupled with mass spectrometry.

Concerning the biological material that has to be examined, due to its complexity, a separation of analytes from the endogenous components of the matrix (e.g., proteins, phospholipids, salts, acids or bases) and impurities (like putrefaction products) is necessary to diminish the load on the chromatographic system. The described methods of the samples’ preparation are diverse and depend on the laboratory. As presented in Table 5, one of the most commonly used preparation technique was precipitation (de Jager et al. 2002; Ma et al. 2007; Tan et al. 2006). Eriksen et al. (2002) and Song et al. (2005) used the procedure of solid-phase extraction an Oasis HLB adsorbent.

In the following two publications (de Jager et al. 2002; Kang et al. 2010) the LOQ of 0.5 µg/L was obtained. In these papers, the authors used precipitation technique, whereas the lowest volume suitable for a successful determination of the analytes was 100 µl (de Jager et al. 2002). On the other hand Kang et al. (2010) used 300 µl of a sample in a two-step procedure including extraction, precipitation, and LLE extraction. In their works, Song et al. (2005) and Ma et al. (2007) described a method that is characterized by LOQ = 1.0 µg/L.

In these cases, 200 μL of biological material were used for extraction, similarly to the herein presented method. The important difference between the methods described in these works and the protocol we propose lies in the use of a different method of sample preparation. Both Eriksen et al. (2002) and Song et al. (2005) used SPE and OASIS, respectively, which is a more time-consuming method than the simple liquid–liquid extraction we applied. LLE is characterized by a short time of preparation, high efficiency, and low operational cost.

An additional advantage of the proposed protocol is the possibility of using it for any type of biological material (including blood, serum, plasma, urine, sections of internal organs and others). Kang et al. in their method managed to achieve the lower LOQ of 0.5 µg/L (Kang et al. 2010). The procedure they developed required a double extraction with acetonitrile, a precipitation and then LLE extraction with hexane–dichloromethane mixtures. In their method, it was necessary to use 300 µl of biological material. The protocol presented in the manuscript allows for the same LOQ as in the work of Eriksen et al. (2002); however, it is enough to extract only 200 μL volume samples. Thanks to the performed optimization, this volume was lowered from 1 mL that was reported by Péhourcq (2004), Kim et al. (2005) and Fortes et al. (2013) in their works.

The undoubted advantages of the presented method are LOQ of 5.0 µg/L and a small volume of biological material needed for extraction. These are the features that, combined with a very simple and fast method of sample preparation, determine a wide application possibilities of the presented method, including both clinical toxicology and forensic analysis.

The proposed analytical methods were successfully applied to a toxicological study. The fully validated method was applied to 28 authentic biological fluids (blood, urine, vitreous humor, bile and stomach content) requiring confirmation for HZ and CZ. The determination of HZ and CZ in vitreous humor, bile and stomach content was performed on urine and blood calibration curves, respectively. The analysis concerned 17 men aged 24 to 72 and 11 women aged 35 to 62. All samples were collected from either: traffic accident (thirteen cases), circulatory and respiratory failure (five cases), multiple drug poisoning (two cases), suicide (one case), death in a psychiatric hospital (one case), sepsis (one case), no clear cause of death (one case), fall from height (one case), lesions in the form of widely developed purulent peritonitis causing respiratory arrest (one case).

Antemortem blood concentrations in one cases (case number 15), CZ concentration was 187.1 ng/mL, but in two cases, antemortem blood (case number 2 and 15) and HZ concentrations were 8.9 ng/mL and 11.2 ng/mL, respectively. In postmortem blood, CZ has been found in concentrations ranging from 4.9 ng/ml to 596.5 ng/ml, and HZ at concentrations ranging from 6.0 ng/ml to 622.0 ng/ml. In different postmortem specimens, concentrations of CZ ranged within 5.9–390.6 ng/mL, 685.4–58,793.5 ng/mL, and 73.1 ng/mL in vitreous humor, urine, and bile, respectively. The concentration of HZ ranged within 4.6–65.7 ng/mL, 61.9–1700 ng/mL, and 47.9 ng/mL in vitreous humor, urine, and bile, respectively. Detailed information regarding determined substances, their concentrations as well as other toxicological findings are gathered in Table 6. In five cases (case number: 6, 7, 9, and 14), toxic level of HZ was found (TIAFT reference blood level list of therapeutic and toxic substances, September 2004. https://www.yumpu.com/en/document/view/13423269/tiaft-reference-blood-level-list-of-therapeutic-and-toxic-gtfch). No subjects died due to the fatal poisoning of CZ or HZ. In six of the cases (case 8, 9, 10, 13, 16, and 22), HZ and CZ were determined only in the blood from the victims or perpetrators of traffic accidents. In 25 of the cases, other xenobiotics were detected including antipsychotics, antidepressants, benzodiazepines, hypnotics and sedatives, and painkillers. Detailed information regarding determined substances, their concentrations as well as other toxicological findings are gathered in Table 6.

Blood HZ concentration exhibited a median of 100 µg/L (range 60–370 µg/L) in 34 person arrested for impaired driving and their averaged 170 µg/L in an additional 5 person (Baselt 2017). In three people whose death was attributed solely to an acute overdose of HZ, its blood concentration was found at the following levels: 0.7, 2.5, and 3 µg/ml. In the case of an 18-year-old woman who died as a result of an overdose of HZ, its concentration in blood and urine was 4.2 and 1.4 µg/ml, respectively (Baselt 2017). The analysis using the developed UHPLC–QqQ-MS/MS method performed on blood, urine, vitreous, and bile samples showed that the HZ concentration was within this range, but lower levels were also observed. It is worth adding that in many of the presented cases, other substances were detected in the biological material, which could have influenced the general condition of the deceased. For example, drugs that depress the central nervous system, i.e., sedatives, hypnotics, anesthetics, benzodiazepines or opioids, act similarly to HZ, which acts synergistically with agents that depress the CNS, thereby intensifying the effects of the latter.