OUP user menu


You-Chen Chao , Shyu-Jye Wang , Heng-Cheng Chu , Wei-Kuo Chang , Tsai-Yuan Hsieh
DOI: http://dx.doi.org/10.1093/alcalc/agg106 431-436 First published online: 12 August 2003


Aims and Methods: Alcoholism may cause a range of diseases including avascular necrosis of the hip joint (AVN), cirrhosis of the liver, pancreatitis and oesophageal carcinoma. Chinese alcoholic patients diagnosed with AVN have a higher incidence of cirrhosis than of acute pancreatitis or oesophageal cancer. Thus, the aim of this study was to investigate genetic differences in polymorphisms of the alcohol-metabolizing enzymes ADH2, ADH3, ALDH2 and P4502E1 for subgroups of Chinese alcoholic patients, defined by diagnoses of AVN (n = 51), acute pancreatitis (n = 92) and liver cirrhosis (n = 159), and for 280 non-alcoholic patients. Results: Analysis revealed that ADH2*1 allele frequency was significantly lower for the alcoholic AVN than for the cirrhosis subgroup. However, no significant difference was found between the alcoholic AVN and pancreatitis subgroups. Furthermore, ALDH2*2 prevalence was not found to differ significantly between the alcoholic subgroups. When compared with our previously published data for alcoholic patients with oesophageal carcinoma, ADH2*1 carriage was significantly less frequent for the alcoholic AVN patients in the current study. Further, ALDH2*2 carriage was significantly less frequent for the alcoholic AVN subgroup than for the oesophageal carcinoma patients. Conclusions: The allele frequencies for ADH2*1 and ALDH2*2 are different when comparing subpopulations of alcoholics defined by presence of specific alcohol-induced diseases, suggesting that genetic variation in alcohol-metabolizing enzyme genes accounts for, at least in part, the specific types of organ damage observed. We also found the combination of AVN and cirrhosis to be more prevalent than that of AVN and acute pancreatitis. In contrast, the ADH2 and ALDH2 allele frequencies for the AVN subgroup were more similar to those of the acute-pancreatitis than to the cirrhosis subgroup. These data indicate the possibility that other genetic variations may also influence the type of organ-specific complications in Chinese alcoholics.

(Received 4 March 2003; first review notified 2 April 2003; in revised form 4 May 2003; accepted 19 May 2003)


It is generally accepted that alcoholism is associated with multiple genetic and environmental factors, and that it is the cause of numerous physical complications (McGue, 1994; Schuckit, 1994, 1995). The reason that certain organ specific complication occurs only in some, but not all, alcoholic subpopulations remains unclear, however.

The first step in the metabolism of alcohol takes place in the hepatocytes, where conversion to acetaldehyde occurs. Acetaldehyde is then metabolised to acetate by aldehyde dehydrogenase (ALDH) (Bosron and Li, 1986; Bosron et al., 1993; Lieber, 1994). The toxic substance, acetaldehyde, and its derivatives are implicated in alcoholic cirrhosis of the liver (Goedde and Agarwal, 1989). The liver alcohol dehydrogenase (ADH), ALDH, and cytochrome P4502E1 (P4502E1) are polymorphic at the ADH2, ADH3, ALDH2 loci, and the 5’-flanking region of the P4502E1. Further, ethnic variations have been reported (Harada et al., 1980; Smith, 1986; Enomoto et al., 1991; Thomasson et al., 1991; Yoshida et al., 1991; Goedde et al., 1992; Kato et al., 1993; Stephens et al., 1994; Chao et al., 1995). The β2β2 enzyme encoded by ADH2 2*2 is approximately 20-fold more active in ethanol oxidation than the β1β1 enzyme (Goedde et al., 1992). Individuals who inherit the ADH2*2 allele have homodimeric and heterodimeric β2-containing isozymes and could be expected to have faster rates of alcohol metabolism and possibly higher concentrations of acetaldehyde production after alcohol consumption. Many studies have attempted to explain susceptibility to alcoholism and alcohol induced liver disease in terms of differences in these alcohol-metabolizing enzymes (Harada et al., 1980; Bosron and Li, 1986; Enomoto et al., 1991; Thomasson et al., 1991; Goedde et al., 1992; Bosron et al., 1993; Chao et al., 1994a; Maezawa et al., 1994; Tsutsumi et al., 1994; Yamauchi et al., 1995). However, the specific mechanisms involved in the different alcohol-induced organ-specific diseases, such as those involving the liver, pancreas, heart and skeletal system, are not yet clear. Maezawa et al.(1996) reported that Japanese alcoholic patients with more severe brain atrophy had a lower incidence of liver cirrhosis. Furthermore, that the ADH2*1 allele frequencies were different for the alcohol induced brain atrophy and cirrhosis subpopulations. These results suggest that subpopulations of alcoholic patients defined by organ specific complications are probably genetically different. In our previous study, it was also demonstrated that Chinese patients with alcohol-induced cirrhosis and acute pancreatitis constituted two genetically distinct subpopulations, with many alcohol induced cirrhosis patients never experiencing acute alcoholic pancreatitis despite the fact that the daily alcohol intake of the former was significantly higher than that of the latter (Chao et al., 1997). Further analysis revealed that the ADH2*1 allele frequency for the pancreatitis patients was significantly lower than for their cirrhotic counterparts. Differences in ADH2*1 and/or ALDH2*1 allele frequencies have also been demonstrated for Chinese alcoholic patients diagnosed with oesophageal cancer in comparison to analogues with acute pancreatitis and cirrhosis (Chao et al., 2000).

Avascular necrosis of the hip joint (AVN) is a specific complication of chronic alcoholism. Clinically, we have noted that this complication was more common for alcoholic patients with liver disease than for those with acute pancreatitis. In this study of alcoholic patients, therefore, we wanted to establish whether the allele frequencies for the alcohol metabolizing enzyme genes for the AVN subgroup are different from those in other alcoholic subgroups and whether they were more similar to the liver cirrhosis subgroup than alcoholic-pancreatitis analogues.


Blood samples were obtained from 302 alcoholic and 280 non-alcoholic patients at the Tri-Service General Hospital in Taipei, from October 2000 to September 2002. Of the 302 alcoholic patients, 51 were diagnosed with avascular necrosis of the hip joint, 159 with cirrhosis and 92 with acute pancreatitis, with all instances of the respective diseases deemed to be alcohol-induced. Of the 280 non-alcoholic patients, 68 were diagnosed with viral hepatitis B and/or C-related cirrhosis [viral cirrhosis: hepatitis B (n = 40), hepatitis C (n = 26), and hepatitis B and C-related (n = 2)], 74 with acute gallstone pancreatitis, 38 with avascular necrosis of the hip joint (non-alcoholic AVN), and 100 were non-alcoholic controls who had been admitted to hospital during the same observation period (clinical diagnoses for the controls included peptic ulcer, inguinal hernia, acute appendicitis, bone fracture and acute gastroenteritis). All of the alcoholic patients had consumed in excess of 60 g alcohol per day, on average, for at least 6 years. None of the patients in the non-alcoholic subgroups had a history of alcoholism and consumption of alcoholic beverages was infrequent for these individuals. All the patients in the alcoholic AVN subgroup (n = 51) had undergone total hip replacement, with diagnosis proven from pathology; for 11 of these, hip-joint involvement was bilateral. Three of these patients (5.9%) had past histories of acute alcoholic pancreatitis, with all also suffering from alcoholic liver disease. One even suffered from oesophageal variceal bleeding; persistent abnormalities in transaminase levels, but no clinical evidence of cirrhosis, were noted for the other two. Eleven of the patients (21.6%) also suffered from alcoholic cirrhosis with oesophageal varices (including the above mentioned acute pancreatitis patient). The other two patients were diagnosed with alcohol-induced liver disease because of persistently abnormal transaminase levels; however, there was no clinical evidence of cirrhosis. Another of these was a case of chronic hepatitis B, with the abnormal liver function possibly a consequence of both diseases. Therefore, at least 15 of the 51 patients (29.4%) had alcoholic liver disease. All the patients without AVN were asymptomatic over hip joints and showed no roentgenographic evidence of AVN. The 159 patients in the alcoholic cirrhosis subgroup were all negative for serum antinuclear and antimitochondrial antibodies, and negative for antibodies to the hepatitis C virus and its RNA (determined using a second-generation test kit, Abbott Laboratories, Chicago, IL, USA) and polymerase chain reaction (PCR), respectively. The hepatitis B surface antigen and HBV DNA (Larzul et al., 1988) were negative for all 159 of these patients. All the cirrhotic patients were graded B or C according to Child–Pugh score (Pugh et al., 1973). These subjects presented with typical sonographic signs suggestive of cirrhosis, and all had endoscopically proven oesophageal varices. Liver biopsy was not performed for most cases because of decompensated hepatic function or massive ascites. None of the patients in the alcoholic cirrhosis subgroup had a history of acute pancreatitis. The acute pancreatitis patients presented with typical symptoms and signs, with elevations of serum amylase and lipase at least three-fold normal levels. Abdominal sonography, computerized tomography, and endoscopic retrograde cholangiopancreatography were used for patient evaluation. For the alcoholic pancreatitis subgroup (n = 92), risk factors for the pancreatitis, other than alcoholism, were carefully assessed before being ruled out. Of these patients, 38 had experienced two or more episodes, with mild but persistent elevations of serum transaminase noted for 14 during a 6-month follow-up period after discharge, indicating the possibility of a coexisting alcohol induced liver disease. The serum albumin, total bilirubin, prothrombin time and peripheral platelet count for these 14 acute alcoholic-pancreatitis patients were within normal limits. No evidence of oesophageal varices or congestive gastropathy was noted for any of the 92 patients in this subgroup. Alcohol consumption histories were obtained using a standard questionnaire. The hospital’s attending physician ensured that the questionnaire was reliably completed by interviewing both the patient and a member of the family, usually the spouse or mother.

ADH2, ADH3 and ALDH2 genotyping and detection of the RsaI and PstI polymorphisms of the P4502E1 gene

DNA was extracted from white blood-cell pellets, obtained after lysis of the red cells with ammonium bicarbonate. The ADH2 and ADH3 genotypes were determined accord-ing to the method of Groppi et al. (1990), with minor modifications (Chao et al., 1994a,b). Briefly, the primers 247 (5′GAAGGGGGGTCACCAGGTTG) and HE45 (5′AATCTTTTCTGAATCTGAACAG) were used for amplification of the ADH2 genes, and the primers 321 (5′GCTTTAAGAGTAAATATTCTGT CCCC) and YC351 (5′AATCTACCTCTTTCCAGAGC) were used for amplification of the ADH3 genes. The PCR-amplification programme for both the ADH2 and ADH3 is set at 35 cycles, with each cycle consisting of 1 min at 95°C, 45 s at 55°C, and 45 s at 72°C.

ALDH2 genotyping was determined by our previously published method, using PCR-directed mutagenesis (Chao et al., 1994a, 1997). The RsaI and PstI polymorphisms of the P4502E1 gene were determined according to the method of Hayashi et al.(1991).


anova, χ2 test and post-hoc multiple comparisons were used for group comparison. The previously published data for ADH2, ADH3 and ALDH2 allele frequencies for alcoholic patients with oesophageal cancer (Chao et al., 2000) were also compared with the analogous data acquired in this study. The experimental values are expressed as mean ± SD. The study design was approved by the National Defense Medical Center Ethics Committee. Informed consent was obtained from all subjects prior to commencement. The characteristics of alcoholic patients in this study are listed in Table 1.

View this table:
Table 1.

Characteristics of alcoholic patients in the different groups in an investigation of alcohol-metabolizing enzyme genes in Chinese alcoholics with avascular necrosis of hip joint, pancreatitis and cirrhosis of the liver


Age, sex and alcohol-consumption details for each group are presented in Table 2. Alcoholic patients with AVN were younger than counterparts with cirrhosis and oesophageal carcinoma, but older than those with acute pancreatitis. The mean average daily alcohol consumption was higher for the alcoholic AVN than for the pancreatitis group. Duration of alcohol consumption history was shorter for alcoholic patients with AVN than for cirrhosis and oesophageal carcinoma patients. The ADH2 genotype and allele frequencies for alcoholic and non-alcoholic patients are presented in Table 3. Carriage of ADH2*1 was significantly less prevalent for the alcoholic AVN than for the oesophageal carcinoma (P < 0.01) and cirrhosis subgroups (P < 0.01), and also significantly less prevalent for the alcoholic pancreatitis than for the oesophageal carcinoma (P < 0.025) and cirrhosis subgroups (P < 0.05).

View this table:
Table 2.

Age, sex and alcohol consumption in the different groups in an investigation of alcohol-metabolizing enzyme genes in Chinese alcoholics with avascular necrosis of hip joint, pancreatitis and cirrhosis of the liver

View this table:
Table 3.

Prevalence of alcohol dehydrogenase ADH2 genotypes in an investigation of alcohol-metabolizing enzyme genes in Chinese alcoholics with avascular necrosis of hip joint, pancreatitis and cirrhosis of the liver

The ALDH2 genotype and allele frequencies for the alcoholic and non-alcoholic subgroups are shown in Table 4. Carriage of ALDH2*2 was significantly less prevalent for the patients with alcoholic AVN than for their counterparts with oesophageal carcinoma (P < 0.001). Furthermore, carriage of ALDH2*2 was significantly more frequent for the oesophageal carcinoma than for the cirrhosis and pancreatitis subgroups; however (with the exception of the oesophageal carcinoma subgroup) significantly more frequent for the alcoholic patients as a whole than for their non-alcoholic counterparts.

View this table:
Table 4.

Prevalence of aldehyde dehydrogenase ALDH2 genotypes in an investigation of alcohol-metabolizing enzyme genes in Chinese alcoholics with avascular necrosis of hip joint, pancreatitis and cirrhosis of the liver

No significant differences were found for ADH3 and P4502E1 genotype distribution and allele frequencies among the studied subgroups (data not shown).

As 16 of the alcoholic AVN subgroup had either been diagnosed with alcoholic liver disease or a history of alcoholic pancreatitis, only the ADH2 and ALDH2 genotypes and allele frequencies for the remaining 35 patients (those with just one complication, alcoholic AVN-1) were used in the comparison with the other alcoholic subgroups (Tables 3,4). Alcoholic patients with both AVN and pancreatitis had more similar ADH2 and ALDH2 allele frequencies than did their counterparts with cirrhosis. The ADH2*1 and ALDH2*2 frequencies were similar for the alcoholic cirrhosis patients and their counterparts with both AVN and cirrhosis.

For further evaluation whether the differences in allele frequency are biologically important, the distribution of ADH2 genotypes in the different alcoholic groups have been compared (Table 3). The results were similar to the allele frequency comparisons.


The pathogenesis of alcohol-related AVN remains unclear. There are multiple theories about the pathogenesis. Two of them are commonly mentioned: (1) thromboemboli in the blood supply to bone resulting from circulating fat, and (2) vessel occluded by excessive packing of the adjacent marrow spaces with extravasated blood or fat (Desforges, 1992; Mont and Hungerford, 1995). For alcoholic Chinese patients in a clinical setting, we noticed that AVN occurs concomitantly with liver disease more frequently than in combination with acute pancreatitis or oesophageal carcinoma. The results of our study — with 29.4% of alcoholic AVN patients also suffering alcoholic liver disease, but only 5.9% of these also experiencing episodes of acute alcoholic pancreatitis — confirms this finding. However, the percentage of alcoholic patients suffering two physical complications was probably higher than in the general population, because our subjects were recruited in a medical centre where they were being treated for severe complications of extended alcohol misuse.

Despite the fact that alcoholic AVN patients had a higher incidence of coexistent alcoholic liver disease, surprisingly the ADH2 and ALDH2 allele frequencies for the alcoholic AVN and acute pancreatitis subgroups were more similar than for the alcoholic AVN and cirrhosis subgroups. The allele frequency of ADH2*1 was only significantly different between patients with alcoholic AVN and cirrhosis of the liver and was not significantly different between patients with alcoholic AVN and pancreatitis. Further analysis showed that the ADH2*1 and ALDH2*2 frequencies were similar for the alcoholic cirrhosis patients and the counterparts with both AVN and cirrhosis (two complications) (Table 4). Analysis of these data suggests two possibilities. First, alcoholic patients with similar ADH2*1 and ALDH2*1 genetic patterns may acquire AVN or pancreatitis. Therefore, in addition to the alcohol-metabolizing enzyme genes we studied, other genetic variations may also influence the type of organ complication in alcoholic patients. An ever-accumulating body of clinical and experimental evidence indicates that inflammatory responses are involved in the pathogenesis of alcoholic liver disease. The association between alcoholic liver disease and polymorphism of the CD14 endotoxin receptor gene, tumour necrosis factor promotor polymorphism and interleukin 10 promotor region polymorphism has been reported (Grove et al., 1997, 2000; Jarvelainen et al., 2001). Second, alcoholic patients with both AVN and cirrhosis (two complications) may be another specific subpopulation that is different from alcoholic patients with AVN only. Alcoholic patients with higher ADH2*1 allele frequency are prone to both AVN and cirrhosis and patients with lower ADH2*1 allele frequency are likely to have AVN only.

By comparing the allele frequencies for the alcohol-metabolizing enzyme genes, we have confirmed that alcoholic patients may be stratified into genetically different subpopulations defined by organ-specific diseases. Previously, we reported that alcoholic patients with acute pancreatitis and oesophageal carcinoma are different from their counterparts with cirrhosis, in terms of ADH2*1 and ALDH2*2 allele frequency. In the present study, we have shown that Chinese alcoholic patients with AVN are different from cirrhotic and oesophageal carcinoma subgroups in terms of ADH2*1 and/or ALDH2*2 allele frequencies. Interestingly, it was also shown that, although the mean age of alcoholic patients with AVN was greater and the daily alcohol consumption significantly higher than for those with pancreatitis, most of the former had never had pancreatitis. These findings suggest that they constituted a distinct subpopulation. Indeed, although statistical significance was not achieved, the ADH2*1 allele frequency was lower for the alcoholic AVN than for the alcoholic pancreatitis subgroup. In fact, the alcoholic AVN subgroup had the lowest ADH2*1 allele frequency of the alcoholic subpopulations studied.

Phenotypic differences between homozygous ADH2 2*2 and ADH2 1*1 is that the former alcoholics have faster ethanol metabolism and consequently may have faster acetaldehyde accumulation, which also depends on the ALDH activity. However, according to the present data, we do not consider acetaldehyde accumulation plays the only major role determining the formation of AVN.

It is well known that tobacco use is also a confounding variable in development of the alcohol-induced organ damage mentioned in this context. In this study, more than 90% of the alcoholics were also smokers. Hence, we cannot clearly answer the question of how the smoking status might impact on our results. In a clinical study it is difficult to control all variables such as drinking pattern, age of onset of alcoholism, sex and coincidence of alcoholic pathologies, which are important in exploring the relationship between polymorphism of ADH and ALDH genes and specific alcohol-induced organ damage.

Although it is now possible to genotype the alcohol-metabolizing enzyme genes and to determine the specific gene polymorphisms responsible for particular inflammatory processes (Grove et al., 1997, 2000; Jarvelainen et al., 2001), the probability of individual complications still cannot be estimated. It seems highly probable that the development of specific complications in alcoholic patients is determined by multiple genes, with most of these still not well understood.

In conclusion, allele frequencies of ADH2*1 and ALDH2*2 appear to differ among disease-defined subpopulations of Chinese alcoholics, suggesting that differences in the alcohol-metabolizing enzyme genes account, at least in part, for the different organ-specific complications. Clinically, alcoholic AVN coexisted more frequently with alcoholic liver disease than with acute alcoholic pancreatitis. However, comparing ADH2 and ALDH2-allele frequencies for the alcohol-induced disease subgroups, the AVN and acute pancreatitis patients were more similar than their AVN and cirrhosis counterparts, suggesting that other genetic polymorphisms may also influence the type of organ-specific complications in chronic alcoholic patients.


This work was supported by grants from the Republic of China’s National Health Research Institute (NHRI-EX90-8939BP) and from the National Science Council (NSC 90–2314-B-016-060, NSC 91–2314-B-016–109).


  • Author to whom correspondence should be addressed at: Division of Gastroenterology, Department of Internal Medicine, Tri-Service General Hospital, No. 325, Section 2, Cheng-Kung Road, Neihu, Taipei, Taiwan, Republic of China. Tel: 886 2 8792 7030; Fax: 886 2 8792 7248; E-mail: chaoycmd{at}ndmctsgh.edu.tw


View Abstract