Olfactory-Related Quality of Life in Multiple Chemical Sensitivity: A Genetic-Acquired Factors Model

Genetic polymorphisms as well as environmental exposures to chemical compounds, iatrogenic, psychological, and physical trauma may play a pathophysiological role in multiple chemical sensitivity (MCS) olfactory complaints, given that xenobiotic metabolism is influenced by sequence variations in genes of metabolizing enzymes. Thus, the aim of the present study was to depict—by means of multiple regression analysis—how different genetic conditions, grouped according to their function as well as clinical background and environmental exposure may interfere with those olfactory complaints referred by MCS patients. Therefore, MCS patients after gene polymorphism sequencing, the olfactory-related quality of life score—calculated by means of the Questionnaire of Olfactory Disorder in forty-six MCS patients—have been found to significantly rely on the phase I and II enzymes score and exposure to previous compounds and surgical treatments. The present work—implementing for the first time a genetic-acquired factors model on a regression analysis—further reinforces those theories, positing MCS as a complex, multifactorial, disease in which the genetic risk related to phase I and II enzymes involved in xenobiotic detoxification, olfactory, and neurodegenerative diseases play a necessary, but probably not sufficient role, along the pathophysiological route of the disease.


Introduction
Multiple chemical sensitivity (MCS) is a relatively common clinical diagnosis in Western populations [1]. The prevalence of self-reported chemical sensitivity symptoms in population-based studies ranges from 9 to 33% [2], whereas physician-diagnosed MCS or reports of disabling consequences in the form of social and occupational disruptions are much lower, ranging from 0.5 to 6.3% [2,3].
MCS patients usually react to a wide range of everyday chemical compounds such as petrol, perfume, or pesticides by complaining of a wide spectrum of symptoms ranging from headache, fatigue, respiratory symptoms, dizziness, nausea, and especially, disosmia [4][5][6]. The discussion on the Legend: Olfactory-related quality of life and clinical-anamnestic aspects in 46 MCS patients. Questionnaire of Olfactory Disorders, QOD; negative statements, NS; positive statements, PS; sum of the scores for the QOD-NS and QOD-PS = quality of life raw score, LQrv. Where needed means ± standards deviations are given.
The analysis of polymorphisms in genes coding for enzymes involved in xenobiotic metabolism pathways showed that the frequencies of CYP2C9*2 and *3 polymorphisms, in heterozygous state and double heterozygous state, were significantly higher in MCS patients compared with those in the Caucasian general population available from the literature ( Table 2). Individuals bearing either the *1/*2 or the *1/*3 genotype were classified as CYP2C9 poor metabolizers, while those with the *2/*3 diplotype were considered very poor metabolizers. Although a higher frequency was observed for CYP2C19 *1/*2 and *1/*17 genotypes, associated with the phenotype poor metabolizer (PM), no significant differences resulted when comparing MCS patients with the Caucasian general population, likely due to the small number of recruited subjects.
The majority of the EM phenotype group possessed the CYP2D6 *1/*1 genotype (28.3%), followed by people having the CYP2D6 *2/*2, and *2A/*2A. Significant differences were found only after comparison of the IM MCS patients with IM in the Caucasian general population (Table 2).
A higher prevalence was found for mutated alleles of GST isoforms, namely GSTP1 313G, GSTM1 Del, and GSTT1 Del as well as for the variant UGT1A1*28 in MCS patients compared with the Caucasian general population. However, these differences did not reach statistical significance, or only in some cases tended to statistical significance, likely due to the small number of recruited subjects.
As a whole, these findings indicate that MCS patients have an impaired metabolism of xenobiotics, both in phase I and phase of body detoxification, and confirm previously reported observations [40,43,44,51,52].
When compared with the Caucasian general population, MCS patients presented with significantly higher frequencies of polymorphisms in genes coding for antioxidant enzymes, namely SOD2 A16V, CAT -C262T, and PON1 L55M (Table 1). Gene polymorphisms examined in our study are known to greatly affect antioxidant enzyme activities due to either resulting amino acid substitutions or the effects on gene transcription rate [53][54][55][56]. The resulting individual phenotype is characterized by a reduced antioxidant defense, potentially leading to increased susceptibility to oxidative stress. The observed increased prevalence of defects in antioxidant enzymes confirms previous results [10,40,51,52] and may provide a mechanistic explanation for reported MCS features of oxidative stress [48].
No significant differences were observed for the distribution of polymorphisms in genes coding for enzymes involved in DNA methylation and repair pathways, namely MTHFR and DNA repair enzyme 8-oxoguanine glycosylase 1 (OGG1) as well as for nitrosative stress-related enzyme NOS3 and immune response enzyme myeloperoxidase (MPO) ( Table 2). The increased frequencies of NOS3 TT894, MTHFR TT677, and MTHFR AC1298 in MCS patients compared with the Caucasian general population are in agreement with previously published observations [40,45,51,57].   Multiple regression analysis was run in order to determine the results of LQrv in relation to the nine prognostic factors. The correlation was statistically significant only for the phase II class score, phase I class score, chemical compounds exposure, and previous surgery with partial correlation coefficient of 0.23, 0.28, 0.3, and 0.2, respectively (Table 3). where X is the predicted value of the LQrv; x 1 is the MTHFR class score; x 2 is the phase II class score; x 3 is the phase I class score; x 4 is the previous event of chemical compounds exposure (0 for absence and 1 for presence); x 5 and x 6 are the previous events of psychological and physical trauma (0 for absence and 1 for presence), respectively; x 7 is the event of previous surgery (0 for absence and 1 for presence); x 8 is the patient's age and x 9 is the patient's gender (1 for female and 0 for male) ( Table 3) for which t-values were contrasted on the significant p-value cut-off on a Pareto chart ( Figure 1). The multiple correlation coefficient was 0.88 with a p value less than 10 −4 . where X is the predicted value of the LQrv; x1 is the MTHFR class score; x2 is the phase II class score; x3 is the phase I class score; x4 is the previous event of chemical compounds exposure (0 for absence and 1 for presence); x5 and x6 are the previous events of psychological and physical trauma (0 for absence and 1 for presence), respectively; x7 is the event of previous surgery (0 for absence and 1 for presence); x8 is the patient's age and x9 is the patient's gender (1 for female and 0 for male) ( Table 3) for which t-values were contrasted on the significant p-value cut-off on a Pareto chart ( Figure 1). The multiple correlation coefficient was 0.88 with a p value less than 10 −4 . Finally, Table 4 and Figure 2 depict the desirability model results in partial and global desirability values for each prognostic factor, particularly depicting a cut-off value of the phase I class score, phase II class score, previous events of chemical compounds exposure, and surgery equal to 1.5, 5.84, 0.52, and 0.3, respectively in order to obtain a LQrv score at least equal to 28.67.  Finally, Table 4 and Figure 2 depict the desirability model results in partial and global desirability values for each prognostic factor, particularly depicting a cut-off value of the phase I class score, phase II class score, previous events of chemical compounds exposure, and surgery equal to 1.5, 5.84, 0.52, and 0.3, respectively in order to obtain a LQrv score at least equal to 28.67.

Discussion
The first interesting aspect of the present study resides in those partial coefficients depicting a multifactorial model contributing to one of the most complained symptoms in MCS, in other words, olfactory alterations which beyond a wide sensorial discomfort referred in this disorder [1,6,12,13,[68][69][70] have been found to severely impact on the quality of life and routine activities of MCS patients [1,6,12,34]. In particular, when paying attention to both the regression model and the Pareto chart highlighting respective t-values accounting for significant impact on LQrv, such a multifactorial model was hierarchically contributed to by genetic factors (phase I and phase II classes scores), environmental (chemical compound exposures), and anamnestic characteristics (presence of previous surgery events) of the patients. This tends to confirm previous hypotheses suggesting the inherited and acquired dysfunction of the chemical defensive system as a molecular basis for MCS complaints [42,48]. Indeed, adequate body response to environmental toxicants presumably requires proper function of the xenobiotic detoxification pathways. Among those factors contributing to variability in human response to toxicants, it can be expected that inherited and acquired variations in the metabolism and excretion of xenobiotics play a major influence. Indeed, it is well known that fat soluble xenobiotics are typically absorbed from the digestive tract, oxidized to intermediates that may be highly reactive, conjugated to increase their solubility, then excreted either by the kidneys in urine, or by the liver in bile [71]. Specifically, elimination of renally conserved, nonpolar fat soluble compounds tends to be more troublesome to the human body and requires sequential metabolic steps. In particular, phase I metabolism of xenobiotics typically activates nonpolar toxic compounds to make them more reactive through oxidation mediated by the cytochrome P450 family of enzymes. On the other hand, phase II conjugation reactions take compounds with an active functional group including compounds activated in Phase I reactions and add an endogenous substrate to that group to make the compounds more soluble and/or reduce their toxicity. Phase II conjugates include glucuronic acid, sulfate, glutathione, acetyl, glycine, and methyl groups [71][72][73]. According to the general trend in literature, oxidative stress and diminished capacity in detoxifying mechanisms have been involved in the pathophysiology of MCS [9]. In particular, the pivotal point in the genesis of the MCS seems to reside in the strict connection existing between oxidative stress damage, neurogenic inflammation, and neural disorders [74,75], based on the vulnerability of the central nervous system to free radicals. In light of this, the present results further strengthen those studies [76,77], highlighting that neurogenic inflammation could be multi-factorially underpinned by genetic predisposition to the breakdown of oxidative stress mechanisms and exposure to environmental events. Following these assumptions, once neurogenic inflammation, together with central biochemical processes [9,78], has been established, according to the neural sensitization theory, MCS could be attributed to a pathological hyper-reactivity of neurons in some areas of the brain, mainly in the olfactory and limbic systems [79], inducing an over-reactivity to external stimuli in different end-organs [6], possibly relying on a olfactory-limbic kindling model [6,79].
Interestingly, the same pathways involved in the metabolism of xenobiotics have been previously linked both to impairment in the olfactory system and involvement in neurodegenerative as well as psychiatric diseases [80]. Previous studies focusing on MCS patients demonstrated altered olfactory sensitivity [1,13,34], although a direct correlation between detoxification pathways, chemical exposures, and/or anamnestic events has not been definitively established, possibly due to weaknesses related to MCS screening, relatively low accuracy of objective olfactory testing, and the absence of correlation between factors and specific tests investigating the quality of life in relation to olfactory disorders. However, many hypotheses have postulated that the chronic exposure (low-dose, overtime) of biogenic amines-based-pesticides (neonicotinoids and formamidines) may disrupt neuronal cholinergic and octopaminergic signaling and produce excessive reactive oxygen species and reactive nitrogen species [9,47,78,81]. These, in turn, may react with macromolecules and interfere with the mitochondrial respiratory chain and mitochondrial Ca2+ metabolism [9,47,78,81]. Oxidative stress has proven to impair cognitive behavior including olfactory learning and memory, especially in those conditions where detoxification pathways may not properly counterbalance the generation of damage [81,82]. This appears to be of much interest, considering that the effects of odors and chemical compounds are not only exerted by the stimulation of the olfactory system through inhalation, but they can also enter the body through absorption in the skin, nose, and mouth, and thereafter, enter the blood stream, thus reaching the brain due to their lipophilic characteristics and causing broad spectrum types of effects [83] such as breakdown in the subjective experience of odor [84].
We here demonstrated significant differences in the distribution of gene polymorphisms of enzymes involved in xenobiotic metabolism, oxidative stress, and DNA methylation/repair pathways between the MCS cohort here studied and the Caucasian general population, with a higher prevalence of gene defects in MCS patients than in healthy subjects ( Table 2). Following the above-mentioned hypothesis that postulates the overlap between breakdown of detoxification pathways, neuropsychiatric disorders, and olfactory dysfunction, it is not surprising that the multiple regression and the desirability model demonstrated that phase I and II scores were increasingly associated with the risk of neuropsychological and quality of life consequences of the olfactory dysfunction in MCS (Table 3, Figure 1). By way of example, GSTs, accounting for phase II class score, are a family of multifunctional enzymes playing a central role in the detoxification of toxic and carcinogenic electrophiles [85], extensively indagated in MCS [52]. In fact, GST isoenzymes catalyze the conjugation of glutathione to a variety of electrophilic compounds including formaldehyde. If the lack of GSTM1 activity, which detoxifies the reactive metabolites of benzo[a]pyrene and other polycyclic aromatic hydrocarbons, is due to homozygous deletion of the gene [86,87], GSTT1 weakness has been found to negatively impact on the metabolism of various potential carcinogens such as monohalomethanes, which are widely used as methylating agents, pesticides, and solvents [86,87]. Furthermore, GST cytosolic activity in olfactory epithelium, the highest among extrahepatic tissues [39,86], is of particular interest in MCS, where the role of odorous triggers is important. In fact, acetaldehyde is one of the most important chemicals that induce sick house syndrome and MCS [88]. As further examples of such behavior, PON1-a high density lipoprotein (HDL)-associated enzyme which reacts with toxic organophosphorus compound including insecticides (malathion) and nerve agents (sarin, somon, and diazinon) [89] by cleaving the homocysteine-thiolactone ring [40,90,91]-is known to be polymorphic in humans, with two isoforms displaying distinct hydrolyzing activities. The Arg192 isoform hydrolyzes paraoxon rapidly, whereas the Gln192 isoform acts slowly [92]. PON1 genes were associated with Gulf War Syndrome whose complaints of olfactory dysfunction are extensively shared with those of MCS [39]. On the other hand, the overlapping symptoms among MCS spectrum disorders have been found to possibly share some common pathogenetic features such as increased nitric oxide/peroxynitrite levels [57] and its consequences on the olfactory pathways [93]. In light of this, SOD2 genetic polymorphisms, which may be considered genetic determinants of MCS risk [40,[42][43][44]52], seem to be connected to the loss of efficiency of detoxification systems, disturbances of free radical/antioxidant homeostasis, and increased production of inflammatory cytokines [31,48,94]. Interestingly, it has also recently been reported that gene variants of NOS2 are associated with the alteration of NO levels in inflammatory bowel disorders, asthma, atopy, olfactory dysfunction, and migraine [57,93,95,96], all of which are comorbidities shared by environmental sensitivity illnesses [57].
In this vision, enzymes also involved in xenobiotic metabolism phase I demonstrated an interface between detoxification function, olfactory perception, and neurodegenerative and psychiatric disorders. In fact, animal models showed that the inhibition of rat CYP P450 monooxygenases increased the electro-olfactogram response amplitude, suggesting a role for these enzymes in signal termination [97], and that an odorant metabolite resulting from the cytochrome-dependent metabolism was able to activate an olfactory receptor [98,99]. CYP2D6, which is present not only in liver, but also at lower levels in brain and other tissues [40,100], metabolizes a wide variety of substances including therapeutic drugs, drugs of abuse, procarcinogens, and neurotoxins [40], and is genetically polymorphic [101]. Thus, the 5-10% of Caucasians who are CYP2D6 homozygous for two non-functional alleles display impaired metabolism of many centrally acting drugs and toxins such as tricyclic antidepressants, selective serotonin re-uptake inhibitors, monoamine oxidase inhibitors, amphetamines, codeine, neuroleptics, and neurotoxins [40]. The enzyme also metabolizes endogenous neurotransmitters [102], which may be related to the observation that poor metabolizers score higher on scales of neuropsychological disorders [103]. Recent literature has provided evidence for a link between the CYP2D6 genotype and many neurodegenerative disorders [40,100,104]. Table 5 summarizes the potential mechanisms involving the above-mentioned genes, their activity, and the polymorphisms affecting their function, possibly underpinning the physiopathological processes of MCS.
However, it is not surprising that the multiple regression and the desirability model highlighted two environmental and anamnestic factors impacting on the LQrv score, worsening the quality of life level of olfactory dysfunction. In particular, the desirability model clearly showed (Table 4, Figure 2)-in line with the Pareto chart built on the regression analysis (Table 3, Figure 1)-that chemical compounds exposure and previous surgery events may accrue, so that the LQrv score may increase over the level of 28.67. Thus, this aspect tends to agree with the literature vision suggesting that a) environmental and medical-surgical exposures may increase the possibility of an over-sensitivity toward external compounds and odorants [31,57] and that b) this may be evidenced in those subjects in which the risk is increased due to a genetic predisposition [31,38,57].
In conclusion, the present work, which implemented for the first time a genetic-acquired factors model on a regression analysis, further reinforces those theories positing MCS as a complex, multifactorial disease in which the genetic risk related to defects in xenobiotic detoxification phase I and II enzymes also involved in olfactory and neurodegenerative diseases plays a necessary, and probably at all not sufficient role in the pathophysiological route for which the combination of multiple aspects including acquired/environmental determinants induce an over-sensitivity to olfactory compounds, finally impacting on quality of life. position of an organic substrate (RH), while the other oxygen atom is reduced to water. This biotransformation reaction is often referred to as "phase I detoxification" in xenobiotic metabolism.

GSTP1 GSTM1 GSTT1
GSH + X  GS-X + H + Conjugation of reduced glutathione (GSH) with toxic agents, carcinogens, drugs, leading to the formation of a detoxified complex more polar and more readily excreted from human body. This reaction is often referred to as "phase II detoxification" in xenobiotic metabolism.  Insertion of one atom of oxygen (monoxygenation) into the aliphatic position of an organic substrate (RH), while the other oxygen atom is reduced to water. This biotransformation reaction is often referred to as "phase I detoxification" in xenobiotic metabolism.
CYP2C9 *2, *3 CYP2C19 *2, *17 CYP2D6 *4, *6, *10, *41 GSH + X  GS-X + H + Conjugation of reduced glutathione (GSH) with toxic agents, carcinogens, drugs, leading to the formation of a detoxified complex more polar and more readily excreted from human body. This reaction is often referred to as "phase II detoxification" in xenobiotic metabolism.  GSH + X  GS-X + H + Conjugation of reduced glutathione (GSH) with toxic agents, carcinogens, drugs, leading to the formation of a detoxified complex more polar and more readily excreted from human body. This reaction is often referred to as "phase II detoxification" in xenobiotic metabolism. Reduction of 5,10-methylenetetrahydrofolate (5,10-MTHF) to 5-methyltetrahydrofolate (5-MTHF), acting as methyl donor for homocysteine (Hcy) remethylation to methionine. Enzyme activity diminishment leads to Hcy accumulation that induces oxidative stress and endothelial dysfunction.

MPO-G463A
Nitric oxide synthase type III NOS3  Insertion of one atom of oxygen (monoxygenation) into the aliphatic position of an organic substrate (RH), while the other oxygen atom is reduced to water. This biotransformation reaction is often referred to as "phase I detoxification" in xenobiotic metabolism.

Participants
We included in the study MCS patients admitted to the Lazio Regional Center for Diagnosis, Prevention and Treatment of MCS and evaluated at "Tor Vergata" University for those symptoms related to smell complaints. Diagnosis of MCS was achieved according to the US Consensus Criteria for MCS [105] and the revisions suggested by Lacour et al. [68,70,106], which were operationalized as follows: (1) symptoms present for at least six months; (2) symptoms occurred in response to exposure to at least two of 11 common volatile chemicals (many of which are reported in Table 6, according to previous studies [107]); (3) co-occurrence of at least one symptom from the CNS and one symptom from another organ system; (4) symptoms cause significant lifestyle changes; (5) symptoms occur when exposed and lessen or resolve when the symptom triggering agent is removed; and (6) symptoms triggered by exposure levels do not induce symptoms in other individuals who are exposed to the same levels. Diagnosis was supported by biochemical analyses, showing high levels of oxidative stress and inflammation (not reported here).
Genomic DNA (gDNA) was isolated from peripheral blood white cells using the PUREGENE-DNA purification system (GENTRA, QIAGEN, Milan), according to the manufacturer's instructions. The gDNA was quantified by spectrophotometric measurement at 260 nm using a Biophotometer (Eppendorf). gDNA quality was considered acceptable for samples with a 260/280 ratio ≥ 1.6. DNA integrity and the presence of contaminant RNA was checked by electrophoresis on 0.8% agarose gel.
The screening for the presence of gene polymorphisms in the above cited genes was carried out by either real-time PCR-based allelic discrimination, direct DNA sequencing, and allele specific PCR.
The PCR reactions were carried out in an Eppendorf Master Cycler Pro ® (Vapo-protect) PCR instrument (Eppendorf, Germany). After 35 cycles of amplification (denaturation at 94 • C for 30 s, annealing at 60 • C for 1 min, and extension at 72 • C for 1 min), the amplification products were electrophoresed in 2% agarose gel, and visualized after staining with ethidium bromide.
PCR products were purified as previously reported [112]. DNA sequencing of PCR products for the MPO gene (289 bp), OGG1 gene (189 bp), and UGT1A1 gene (351 bp) was performed using the Sanger method employing the BigDye Terminator v1.1 Cycle Sequencing kit (Applied Biosystems, Life Technologies, Milan, Italy). The reaction was carried out in a final volume of 20 µL, containing 25 ng of an appropriate amount (20 ng/100 bp) of purified PCR product, 5 pmol of the relevant primer, 2.5_Ready Reaction mix, 1_Sequencing Buffer, and DNAse/RNAse-free water, in the Eppendorf Master Cycler Pro ® (Vapo-protect) PCR instrument (Eppendorf, Germany). The thermal cycling conditions were: 96 • C for 30 s and 28 cycles, 30 s at 96 • C, 10 s at 50 • C, 4 min at 60 • C, and finally 5 min at 4 • C. The cycle sequencing products were purified and analyzed on an automated ABI PRISMs 310 Genetic Analyzer (Applied Biosystems, Applera Corp., Milan, Italy) as previously reported [112].
Genotyping calls were manually inspected and verified prior to release. MPO and OGG1 mutated genotypes were assigned on the basis of single nucleotide substitutions, when present. UGT1A1 genotypes were assigned based on the number of TA repeats for each allele (i.e., 6/6, 6/7, 7/7, or 7/8).

Multiplex PCR Analysis for Deletion Polymorphisms
The deletion polymorphisms for the GSTM1 and the GSTT1 genes were determined simultaneously in a single assay using a multiplex PCR approach with the amplification of the GSTM1 and the GSTT1 genes from genomic DNA and using β-globin as the internal control, as previously described [44].

Olfactory Study
A detailed case history investigating events beyond the education level and occupation of chemical compounds exposure, physical and psychological trauma, and previous surgery was performed in all MCS subjects who underwent ear-nose-throat examination with fiberoptic check of the upper airways.
For the study of olfactory complaints during daily activities, we used the "life quality statements" from QOD that expressed the patients' complaints related to the smelling difficulties [1,12,113]. There were in total nineteen life quality statements: seventeen negative statements (QOD-NS) with 3 points assigned for checking the box "agree", 2 points for "partly agree", 1 point for "partly disagree", and 0 points for "disagree", and two positive statements (QOD-PS) with 0 points by checking the box "agree", 1 point "partly agree", 2 points "partly disagree" and 3 points "disagree". The sum of the scores for the QOD-NS and QOD-PS produces the LQ raw score (LQrv); a maximum of 57 points can be reached. High scores indicate a strong impairment [113].
Subjects with diabetes, oncologic or HIV history, neurological, and psychiatric or mood disorders, radiation, and traumatic brain injury were excluded from the study. No patient showed liver or renal abnormalities nor was pregnant or breastfeeding. Neurological diseases were excluded with the mini mental state examination and magnetic resonance imaging (MRI). All conditions that could potentially develop an olfactory dysfunction were considered as exclusion criteria. Thus, patients with sino-nasal disorders; neuro-psychiatric disorders (Parkinson's disease, Alzheimer's disease, schizophrenia, multiple sclerosis, and depression); lower airways and/or lung diseases; active hepatitis, cirrhosis, chronic renal failure, vitamin B12 deficiency; alcohol, tobacco, or drug abuse; cerebral vascular accidents; insulin dependent diabetes mellitus; hypothyroidism; and Cushing syndrome were not included in the study [114].

Data Handling and Statistical Analysis
In order to carry out statistical analyses, all selected polymorphisms were sub-grouped according to their functional effects and the enzyme role [71,115]. The following three classes were obtained: MTHFR (including MTHFR C677T and MTHFR A1298C), phase I (including CYP2D6, CYP2C19, UGT1A1*28, CYP2C9 A1075T, and CYP2C9 C430T), and phase II (including MPO G463A, GSTP1 A313G, GSTM1, GSTT1, SOD2 RS4880, CAT C262T, OGG1 C315G, PON1 A575G, PON1 C108T, and eNOSAsp298Glu). Wild-type, heterozygote, and mutated homozygote genotypes of the MTHFR C677T, MTHFR A1298C, MPO G463A, GSTP1 A313G, SOD2 RS4880, CAT C262T, OGG1 C315G, PON1 A575G, PON1 C108T, eNOSAsp298Glu, UGT1A1*28, CYP2C9 A1075T, and CYP2C9 C430T were vectorially scored from 0 to 2, respectively. Wild-type and deleted genotype of both GSTM1 and GSTT1 were scored 0 and 1, respectively, and the ultra-rapid, extensive, intermediate and slow metabolizer phenotype of CYP2D6 and CYP2C19 was scored from −1 to 2. A composite score for each class was computed by adding the score of the single form of the enzymes. Absence and presence of events of chemical compounds exposure, physical and psychological trauma, and previous surgery were scored 0 and 1, respectively. Number and percentage of cases for each gene condition and clinical-anamnestic aspect was calculated and compared with the literature findings by means of the Fisher's exact test. The X 2 test was carried out to define associations between categorical factors and groups. To assess that data were of Gaussian distribution, D'Agostino K squared normality and Levene's homoscedasticity test were applied (where the null hypothesis is that the data are normally and homogenously distributed) [116].
Given the exploratory nature of the study and previous biomedical approaches [117], correlations between the LQrv and nine prognostic factors (genotypic/phenotypic score of MTHFR, phase I and phase II classes, absence/presence of events of chemical compounds exposure, physical and psychological trauma and previous surgery, age, and gender (0 = male, 1 = female) were examined by a multiple regression analysis. P-values less than 0.05 were considered as statistically significant. A Pareto chart contrasting the t-values and p-values of each factor were generated. Finally, a two-sided desirability model was achieved for each prognostic factor, according to the mentioned regression model and setting upper and lower limit of LQrv at 22 and 40, representing the lowest and highest desirability values, respectively [117,118]. Subsequently, the partial desirability functions (d i ) were combined into a single composite global desirability function, defined as the geometric mean of the different d i values, implying that all responses are in a desirable range simultaneously and the combination of the different criteria is therefore globally optimum, so the response values are near the target values.

Limitations of the Study
The results of this study have to be considered with caution, because of their preliminary nature and considering the following limitations. In fact, no correction for multiple comparisons was performed in the multiple regression model. If this exploratory-rather than confirmatory-approach is extensively adopted in similar studies [117,119,120] to reduce the likelihood of Type II statistical errors as much as possible, this choice may have induced the increase in likelihood of Type I statistical errors. However, given these assumptions, the present data have to be considered as preliminary and should be replicated in further cohorts of patients.