Dextrorphan

We report the results of an investigation conducted in 992 healthy control subjects (854 adults and 138 adolescents) and in 116 subjects with lung cancer (LC) for the purpose of detecting those individuals with a possible genetic predisposition to lung cancer. The test consists of the oral administration of 64 [micro]mol of dextromethorphan (DMP) with collection of urine samples over the following 8-h period and urine assay of the drug (DMP) and its main metabolite, dextrophan (DOP). The ratio of the urinary concentrations of DMP to those of DOP is called the metabolic ratio (DMP/DOP) and is inversely proportional to the DMP demethylation rate. The pattern of the metabolic ratio ([Log.sub.10] DMP/DOP) allowed, using a maximum likelihood approach, the identification of three subpopulations in the 854 control subjects (adults): (1) probable homozygous extensive metabolizers with [Log.sub.10] DMP/DOP [less than] - 1.74 (73.1 percent); (2) probable heterozygous intermediate metabolizers with [Log.sub.10] DMP/DOP in the -- 1.74 to -- 0.40 range (22.3 percent); and (3) probable homozygous poor metabolizers with [Log.sub.10] DMP/DOP [greater than] -- 0.4 (4.6 percent). Most of the patients with LC (89 percent) were probable homozygous extensive metabolizers. As the latter have a cancer risk that is 2.54-fold greater than that of intermediate metabolizers (95 percent confidence interval [CI]: 1.37 to 4.73) and 7.43-fold greater than that of poor metabolizers (95 percent CI: 1.01 to 54.5), their identification by means of the DMP test may be particularly useful for subjects exposed to environmental and occupational carcinogens. The phenotype test used is similar to that of the debrisoquin test, but presents the advantage that DMP is a widely used, harmless drug with a faster and simpler urinary assay procedure.

The enzyme oxidative systems involved in metabolism and in the subsequent activation of environmental carcinogens, such as the hydrocarbons of cigarette smoke and the combustion products of petrol and diesel oil, are thought to be the same as those utilized in the metabolism of numerous pharmacologically active substances such as sparteine and the antihypertensive agent, debrisoquin.[1-5]

The ratio of the urinary concentration of the drug to that of its respective metabolite is called the metabolic ratio and is inversely proportional to the metabolization rate.

The debrisoquin/hydroxydebrisoquin metabolic ratio presents a trimodal distribution (extensive, intermediate, and poor metabolizers), presumably related to the three phenotype EM, IM, and PM, respectively. This is related to individual ability to metabolize greater or lesser amounts of drug in the given time unit.[4,5]

Research studies conducted using the debrisoquin test in healthy subjects and in subjects with lung cancer (LC) have revealed a distinct increase in cancer risk in homozygous extensive metabolizers, presumably due to a greater concentration of "activated" environmental oncogenes in the tissues,[6-8] even if some authors diminished the importance of these findings.[9,10]

A number of investigators[11] have demonstrated, on the basis of comparison of the respective metabolic ratios, a close correlation between the o-demethylation of dextromethorphan (DMP) and the 4-hydroxylation of debrisoquin. It would therefore seem appropriate to use DMP, which is a widely used, harmless drug, as a substitute for debrisoquin for the purposes of studying genetic differences in oxidative metabolism[11,12] and thus of identifying homozygous extensive metabolizers who are more susceptible to cancer risk due to exposure to cigarette smoke and/or environmental exposure to combustion products.[13,14]

To this end we decided to carry out an investigation of the case-control type based on administration of DMP and on assay of its urinary metabolite, dextrorphan (DOP). The investigation was conducted in 992 healthy control subjects (854 adults and 138 adolescent students) and in 116 subjects with LC.

METHODS

Determination of urinary concentrations of DMP and its metabolite, DOP, was performed in a group of 116 consecutive subjects with LC admitted to the Surgery Clinical of Verona University Hospital in the period January 10, 1988 to January 3, 1990; diagnosis was histologically ascertained by biopsy specimens or by histologic analysis of surgical specimens. These subjects, whose ages ranged between 45 and 78 years, had no major concurrent diseases, particularly liver or kidney lesions or other tumors, and they were not receiving any form of pharmacologic treatment. Subjects with a medical history of possible occupational exposure or with stage M1 tumors were excluded. All subjects were studied before surgery.

[TABULAR DATA OMITTED]

The control group consisted of a sample of healthy subjects drawn from a population of 6,000 workers in an Italian telephone industry. Forty-six of the 900 subjects sampled refused to participate in this study, so the test was performed in 854 adults (aged between 25 and 65 years). None of them had a history of alcohol or drug abuse, and none were receiving treatment with quinidine or neuroleptics. The test was performed also on a group of 138 adolescents attending school and ranging in age from 14 to 18 years.

The test was based on the oral administration of three 7.5-mg tablets of dextromethorphan bromohydrate (Bronchenolo Tosse, Midy, Italy), corresponding to a total of 64 [micro]mol, at bedtime after emptying the bladder. Subjects drank no alcohol during the test or in the 4 h prior to administration of the drug. All urine passed over the 8-h period following administration was collected, and the samples were used for determination of DMP and DOP concentrations using the method described in a previous study.[15] Individual differences in DMP metabolism are expressed as [Log.sub.10] of the metabolic ratio of the urinary concentration of DMP to that of its metabolite. All the tests were performed between January 10, 1988 and January 3, 1990.

Statistical Analysis

Comparisons between groups were performed using nonparametric tests (the Mann-Whitney U test and the Kruskal-Wallis W test[16]). Comparisons among proportions were performed using the [[chi].sup.2] test using a statistical package (Stat-Xact); also "exact" 95 percent confidence intervals (CIs) for the odds ratios were calculated using this software.

The overall distribution of the decimal logarithm of the metabolic ratio was assumed to be a mixture of three normal distributions, one for each hypothesized phenotypic group: probable homozygous dominant extensive metabolizers (phenotype EM), probable heterozygous intermediate metabolizers (phenotype IM), and probable homozygous recessive poor metabolizers (phenotype PM). Maximum likelihood estimates of the nine parameters of the overall distribution (ie, the three means and the three SDs of the three normals and the proportion of the total in each normal component) were obtained using Newton methods and the MaxLik procedure included in the GAUSS Application Modules.[17]

[TABULAR DATA OMITTED]

Likelihood ratio tests were performed to evaluate consistency of the theoretical proportions of the three genotypes with those expected under Hardy-Weinberg equilibrium. Cutpoints for the categorization of metabolic ratio (MR) into phenotypic groups were determined by calculating the values of log MR where the adjacent normal densities were equal.

RESULTS

DOP and DMP Excretion

Table 1 shows descriptive statistics of urinary DOP, DMP, MR, and its [Log.sub.10] in the control subjects studied. Total 8-h urinary excretion of the metabolite, DOP, ranged from 0.2 to 63 [micro]mol (with a median of 17.3, a mean of 18.9, and a SD of 12.2). The amount recovered as DOP ranged from 0.3 to 99 percent of the drug dose administered. The amount of the drug DMP found in the urine samples of the same subjects ranged from 0.013 to 20.8 [micro]mol (with a median of 0.170, a mean of 0.481, and a SD of 1.273). When the nonparametric U and W tests were used, no significant differences in the urinary excretion of DOP and DMP in relation to age, sex, or smoking habits were found.

Determination of Type of Metabolizer

In the 854 control subjects (adults), the distribution of MR values, which ranged between 0.001 and 14.41, was markedly skewed to the right with a skewness coefficient of 8.3. The median of the distribution was 0.010, while the mean was 0.196 and the SD 1.052. Also, the distribution of the decimal logarithm was skewed to the right, with a median of -- 1.998, a mean of -- 1.827, and a SD of 0.667. The diagram of the frequencies of the log MR is reported in Figure 1; there is a long right tail that could be due to the coexistence of other distributions in addition to the principal one.

These data were then analyzed, using maximum likelihood method, in order to fit a three normal mixed model, corresponding to phenotype EM, IM, and PM. The results are shown in Table 2, while in Figure 1, the predicted distribution of log MR superimposed on the observed one in the entire group of control subjects is depicted.

The EM subjects had a log MR mean of -- 2.093 with a SD of 0.243; IM subjects had a log MR mean of -- 1.406 with a SD of 0.598; PM subjects had a log MR mean of 0.380 with a SD of 0.415. The EM group was about 73 percent of the entire control population, while IM contributed 22 percent, and PM the remaining 5 percent.

[TABULAR DATA OMITTED]

The predicted distributions of log MR by phenotypic groups are also displayed in Figure 1. The maximum separation between extensive and intermediate metabolizers was achieved setting a cutoff at log MR = -- 1.740, corresponding to MR = 0.176, while the "best" cutoff between intermediate and poor metabolizers was found to be log MR = -- 0.4014, corresponding to MR = 0.669.

Table 3 summarizes the results of DMP and DOP analysisin the urine of EM, IM, and PM phenotypes identified using the previously mentioned cutoffs. Also in this case when the nonparametric U and W tests were used, no significant differences in the urinary excretion of DOP and DMP among phenotypic groups in relation to age, sex, or smoking habits were found.

When a model constrained to obey to Hardy-Weinberg equilibrium was fitted to the control subjects data, the estimates we obtained were quite similar to those found in the unconstrained case (Table 4).

Despite the striking similarity, the likelihood ratio test resulted in a significant difference between the two models ([[chi].sup.2] = 7.14, p [less than] 0.05). These results, in our opinion, could be due to the very large sample size.

DOP and DMP Excretion in Patients With LC

In the 116 patients with LC, with a mean age of 62 [+ or -] 8 years, the mean total 8-h urinary excretion of DOP was 19.2 [+ or -] 15.5 [micro]mol. The amount of drug excreted ranged from 0.04 to 6.6 [micro]mol with a mean value of 0.22 [+ or -] 1.0. The MR values ranged from 0.001 to 0.741 with a mean value of 0.023 [+ or -] 0.079 and a median of 0.007 [micro]mol, while the [Log.sub.10] transformed values had a median of - 2.180, a mean of -- 2.107, and a SD of 0.487.

Case-Control Comparison

From the data reported above, it is evident that the urinary excretion of the metabolite, DOP, is greater in patients with LC than in control subjects, while the amount of excreted drug is less in the LC group. The MR and log MR are therefore higher in control subjects than in patients with LC. Nonparametric U test confirmed that all these differences were highly significant (MR: U = 5.14, p [less than] 0.001; [Log.sup.10] MR: U = [greater than] 4.66, p [less than] 0.001). When the previously mentioned cutpoints were used to divide cases and control subjects in the three phenotypic groups, the results shown in Table 5 were obtained.

The prevalence of LC (14.2 percent, 6.1 percent, 2.2 percent) is significantly different in the three phenotypic groups ([[chi].sup.2] = 13.99; p = 0.0012). Considering PM as the referent group, the crude odds ratio for LC risk is 2.92 for IM (CI, 0.37 to 23.05) and 7.43 for EM (95 percent CI, 1.01 to 54.50); if we consider IM as the referent group, the crude odds ratio for LC risk is 2.54 for EM (95 percent CI, 1.37 to 4.73). So, an increasing risk for LC was observed across the three phenotypic groups, with the highest risk for EM and the lowest for PM.

DISCUSSION

The data reported in this study show that the odemethylation rate, as expressed by the urinary drug/metabolite ratio, presents a trimodal distribution pattern in a sample of 854 healthy control subjects.

Phenotype determination, based on the metabolic ratio, PM ([greater than] 0.669), IM (0.176 to 0.669), and EM ([less than] 0.176) reasonably reflects the genotype distribution. The situation is that probable homozygous EM amounts to 73.1 percent of the population, probable heterozygous IM to 22.3 percent, and PM to 4.6 percent. This distribution is in close agreement with the result of Caporaso et al.[7] Whereas subjects with high MR (PM) are very likely to be homozygous genotype, the distinction between probable heterozygous intermediate metabolizers (IM) and probable homozygous extensive metabolizers (EM) is possible with the aid of family pedigree analysis data.

In the three subpopulations, EM, IM, and PM, no significant differences due to age, sex, or smoking habits were noted.

Figure 2 compares the semilogarithmic scatter diagrams for the distribution of the frequencies of log MRs in healthy control subjects (top panel) and in patients with LC (bottom panel). It may be noted that the MR values in the patients with LC are grouped toward the left of the histogram in the homozygous extensive metabolizer area. On establishing the cutoff between IM and EM at the -- 1.740 value, as mentioned above, the result is that probable homozygous EM patients make up the vast majority of the LC population (89 percent). Among the patients with LC, in fact, only one was a poor metabolizer, and it should be noted that this patient smoked 30 cigarettes a day histologic subtypes least correlated with the EM phenotype. The data show that the MR in patients with LC is on average about ten times lower than in normal control subjects.

The question arises as to whether the significant differences between MRs in control subjects and patients with LC are the cause or the result of the LC. Several data support the former hypothesis. Studies in animals have shown mono-oxygenase activity to increase in the presence of tumor.[1] This leads to an increase in MRs, and thus the presence of tumor lesions would bring about a decrease rather than a rise in the percentage of poor metabolizers. In a previous study of ours,[18] we reported the results of an investigation in 35 smokers with urinary bladder cancer. The distribution of DMP MRs in these subjects showed no significant difference from that in healthy control subjects. It has also been observed that oxidative metabolism determinations such as the ratio of mephenytoin to plasma clearance of antipyrine remained unchanged in a small sample of patients with LC with a prevalence of extensive metabolizers.[19]

It can be assumed, then, in agreement with Nebert's findings for debrisoquin metabolism,[20] that the influence of environmental or nongenetic factors (cigarette smoke, age, sex, alcohol intake, and concomitant disease) on metabolic effects is insufficient to account for effects of the magnitude of the genetic effects in the case of DMP metabolism (extensive metabolizers).

If EM subjects are exposed to two or more risk factors, a synergistic effect is often produced. Caporaso et al[7] observed that, in subjects who were smokers and extensive debrisoquin metabolizers, and whose relative risk compared with PM was 4.0, there was a 35 percent increased risk among those who were exposed to occupational carcinogens of the polycyclic aromatic hydrocarbons (PAH) type. Our data, which show a higher risk in the EM phenotype compared with the PM and IM phenotypes, provide indirect proof of the role of an enzyme inducing activation of DMP and of a carcinogen present in cigarette smoke and in the combustion products of petrol and diesel oil. We may therefore assume that the absence of apparent sources of error prompts us to postulate a genetic predisposition to LC.

The results of the present study show that subjects who are both extensive DMP metabolizers and cigarette smokers are at high risk of LC. Subjects who, in addition to this, are exposed to occupational carcinogens of the PAH type are at even greater risk. For practical purposes, the DMP phenotyping protocol is the same as that for debrisoquin, except that the DMP and DOP analysis is faster and simpler. It therefore proves possible to employ DMP, which is a more extensively used and harmless drug, to establish genetic differences of oxidative metabolism of this type and thus to identify extensive metabolizers who are at high risk of LC. This can be done particularly for certain overexposed categories such as young smokers, traffic police, motorway toll gate attendants, etc.

REFERENCES

[1] Miller JA, Miller EC. The metabolic activation and nucleic acid adducts of naturally occurring carcinogens; recent results with ethy carbamate and the spice flavours safrole and estragole. Br J Cancer 1983; 489:1-15

[2] Steiner E, Iselius L, Alvan G, Lindsten J, Sjoqvist F. A family study of enetic and environmental factors determining polymorphic hydroxylation of debrisoquin. Clin Pharmacol Ther 1985; 38:394-400

[3] Eichelbaum M, Baur MP, Dengler HJ, Osikowska-Evers BO, Tieves G, Zekorn C, et al. Chromosomal assignment of human chromosome P-450 (debrisoquine/sparteine type) to chromosome 22, Br J Clin Pharmacol 1987; 23:455-58

[4] Skoda RC, Gonzalez FJ, Demierre A, Meyer UA. Two mutant alleles of the human cytochrome P-450dbl gene (P450C2DI) associated with genetically deficient metabolism of dibrisoquine and other drugs. Proc Natl Acad Sci USA 1988; 85:5240-43

[5] Gonzalez FJ, Skoda RC, Kimura F, Umeno M, Zenger VM, Nebert DW, et al. Characterization of the common genetic defect in humans deficient in debrisoquine metabolism. Nature 1988; 331:442-46

[6] Ayesh R. Idle Jr, Ritchie JC, Crothers MJ, Hetzel MR, Metabolic oxidation phenotypes as markers for susceptibility to lung cancer. Nature 1984; 312:169-70

[7] Caporaso N, Tucker MA, Hover RN, Hayes RB, Pickle LW, Issaq HJ, et al. Lung cancer and the debrisoquine metabolic phenotype. J Natl Cancer Inst 1990; 82:1264-722

[8] Law MR. Review: genetic predisposition to lung cancer. Br J Cancer 1990; 61:195-206

[9] Roots I, Drakoulis N, Loddenkemper R, Minks T, Nitz M, et al. Debrisoquine hydroxylation phenotype, acetylation phenotype, and ABO blood groups as genetic host factors of lung cancer risk. Klin Wochenschr 1988; 66:87-97

[10] Benitez J. Lodero JM, Jara C, Carrillo JA, Cobaleola J, Llerena A, et al. Polymorphic oxidation of debrisoquine in lung cancer patients. Eur J Cancer 1991; 27:158-61

[11] Schmid B, Bircher J, Preisig R, Kupfer A. Polymorphic dextromethorphan metabolism: co-segregation of oxidative O-demethylation with debrisoquin hydroxylation. Clin Parmacol Ther 1985; 38:618-24

[12] Hildebrand M, Seifert W, Reichenberger A. Determination of dextromethorphan metabolizer phenotype in healthy volunteers. Eur J Clin Pharmacol 1989; 36:315-18

[13] Faccini GB, Bertozzo L, Zatti N. Determination of dextromethorphan in urine by gas chromatography. In: Abstracts of the 15th World Congress of Anatomic and Clinical Pathology. Florence, Italy, 1989; 165

[14] Faccini GB, Puchetti V, Zatti N. Dextromethorphan oxidation phenotypes as markers for susceptibility to lung cancer. Clin Chem 1990; 36:387

[15] Zatti M, Faccini GB, Pasini F, Bertozzo L, Perazzoli P. A new gas chromatographic method for identification of subjects who are genetically at risk of lung cancer [abstract]. Presented at the 23rd Congress of Italian Society of Clinical Biochemistry, Grado, September 15-18, 1991. Biochim Clin 1991; 15(Suppl 2/10):71

[16] Siegel S. Non-parametric statistics: for the behavioral sciences. New York: McGraw-Hill Book Co, 1956

[17] GAUSS Application Modules, System Version 2.01. Washington, DC: Aptech Systems Inc, 1991

[18] Faccini GB, Franzolin N, Monaco C, Comunale L, Zatti N, Mobilio G. Observations about genetic risk in bladder cancer. Acta Urol Ital 1990; 6:345-49

[19] Ayesh R, Idle JR. Evaluation of drug oxidation phenotypes in the biochemical epidemiology of lung cancer risk. In: Boobis AR, Caldwell J, De Matteis F, Elcombe CR, eds. Proceedings of the Sixth International Symposium Microsomes and Drug Oxidation. London: Taylor and Francis, 1985;340-86

[20] Nebert DW, Possible clinical importance of genetic differences in drug metabolism. BMJ 1981; 283:537-42

COPYRIGHT 1994 American College of Chest Physicians
COPYRIGHT 2004 Gale Group