Laboratory testing for the antibodies that cause heparin-induced thrombocytopenia: how much class do we need?

Heparin-induced thrombocytopenia (HIT) is usually caused by platelet-activating antibodies of immunoglobulin G class that recognize platelet factor-4 (PF4) bound to heparin or certain other polyanions. Commercial enzyme immunoassays (EIAs) for PF4/polyanion-reactive antibodies detect two immunoglobulin classes (IgA and IgM) besides IgG. To investigate whether the additional detection of these antibody classes improves or worsens assay operating characteristics, we compared the sensitivity and specificity of EIAs that detect these 3 immunoglobulin classes individually with that of a commercial EIA (Genetic Testing Institute, GTI), as well as a platelet-activation assay, the serotonin-release assay (SRA). We compared the operating characteristics of these 5 assays by evaluating 448 patients, in 14 of whom clinical HIT developed, who received either unfractionated or low molecular weight heparin in prospective studies that included systematic platelet-count monitoring and serologic evaluation for anti-PF4/polyanion antibodies. We found that the SRA and IgG and commercial EIAs had similar high sensitivity for HIT; however, diagnostic specificity (for unfractionated and low molecular weight heparin, respectively) varied considerably, as follows: SRA (95.1%, 97.2%) > IgG EIA (89.0%, 93.7%) > GTI EIA (74.2%, 87.6%). Additional detection of IgA and IgM antibodies by the GTI EIA worsened test specificity by detecting numerous nonpathogenic antibodies. Moreover, the frequency and magnitude of IgA and IgM antibody formation in non-HIT immune responses did not differ from that exhibited by patients in whom clinical HIT developed. We conclude that an EIA that detects anti-PF4/polyanion antibodies of only the IgG class has greater diagnostic usefulness in revealing clinical HIT than does an assay that also detects IgA and IgM class antibodies.     

Eliminating interference from heterophilic antibodies in a two-site immunoassay for creatine kinase MB by using F(ab')2 conjugate and polyclonal mouse IgG.

Heterophilic antibodies interfere in two-site immunoassays. Our purpose was to screen for the samples containing heterophilic antibodies that react with mouse immunoglobulins and to use them to determine how to eliminate interference in various two-site immunoassays being developed in our laboratory. Of approximately 2600 samples screened, 81 had heterophilic antibodies. When creatine kinase MB (CKMB) concentration was measured with intact antibody conjugate in these 81 samples, 18 (22%) samples had apparent CKMB values significantly greater than values measured with Hybritech's "Tandem -E CKMB immunoenzymetric" assay and Corning agarose-gel electrophoresis. Adding up to 133 mg/L of polymerized IgG or up to 1666 mg/L polyclonal mouse IgG in the assay did not eliminate the interference in all the samples. However, adding F(ab')2 conjugate plus 40 mg/L of polymerized IgG or 83 mg/L of polyclonal mouse IgG eliminated the interference in all the samples. This approach was also effective in eliminating the interference in 15 samples containing 4.7-165.2 mg/L of human antimouse antibody (HAMA). Combined use of F(ab')2 conjugate and polyclonal mouse IgG is recommended to eliminate interference from heterophilic antibodies that react with murine immunoglobulins or HAMA in two-site murine-antibody-based assays.     

Direct quantitation of IgG subclasses 1, 2, and 3 bound to red cells by Rh1 (D) antibodies

The authors developed quantitative radioimmunoassays to allow direct measurement of total human IgG and individual IgG subclasses among antibodies bound to cell surfaces. The assays use four mouse monoclonal radioiodinated antibodies, one that reacts equally well with all four human IgG subclasses and three that are specific for human IgG subclasses 1, 2, or 3. The assays were used to analyze IgG subclass composition in 21 high-titer anti-D samples from Rh-negative volunteers immunized for Rh immunoglobulin production. Anti-D activity was restricted primarily to the IgG1 and IgG3 subclasses. Eleven of 21 sera demonstrated red cell antibodies with a marked predominance of IgG1 (87 +/- 3.6% of total IgG antibody, +/- SEM) and low levels of IgG3 (1.4 +/- 0.73%). In the remaining 10 sera, IgG3 made up a greater proportion of total IgG antibody (32 +/- 3.8%), although IgG1 was still predominant (61 +/- 4.1%). This observed dichotomy in the IgG subclass profiles of different anti-D sera may be a consideration in the selection of anti-D sera for the production of the immunoglobulin used in the prophylaxis of Rh-incompatible pregnancies.     

Class- and subclass-specific antibody response to hemocyanin in nonmalignant paraproteinemia.

IgM, IgG, and IgA class-specific, as well as IgG subclass-specific antibody titers against the primary immunogen HPH were measured with ELISA in 19 patients with nonmalignant paraproteinemia (eight with IgG1, two with IgG2, two with IgG4, four with IgM, and three with IgA) and in a simultaneously studied age- and sex-matched control group. After primary immunization only IgM and IgA anti-HPH titers were significantly lower in the patient group. Four patients with relatively high IgG or IgA serum paraprotein levels did not produce antibodies in some Ig classes or IgG subclasses, whereas all other patients and all controls developed antibody titers in all classes and IgG subclasses. Low or absent antibody titers did not occur preferentially in the Ig (sub)classes to which the paraproteins belonged. After secondary immunization the patients could not increase or maintain their antibody titers as well as the controls, and this was most clear in the IgM and IgA antibody class. A direct correlation between polyclonal serum IgM levels and IgM anti-HPH titers was present in the patients. Such a correlation was absent for IgA in the patients and for all classes in the controls. It is concluded that humoral immunosuppression as measured with a newly encountered antigen in patients with nonmalignant paraproteinemia is most clearly expressed in the IgM and IgA antibody class and that the paraprotein (sub)class is not preferentially involved.     

Studies of possible antiidiotypic control mechanisms in Graves' disease

Serum from 55 patients with active Graves' disease and 55 patients who had received successful treatment (in whom the disease was inactive) were examined for the presence of possible antiidiotypic antibodies with an enzyme-linked immunosorbent assay (ELISA) for anti-F(ab')2. Murine IgG monoclonal antibodies (Mabs) against human thyroid-stimulating hormone (TSH) and human TSH receptors were also used as antigens in parallel ELISA assays. Patients with active and patients with inactive Graves' disease showed elevations of IgG anti-F(ab')2 antibodies when compared with normal controls. Similarly, both active and inactive Graves' disease sera showed higher levels of IgG anti-LE4, a mouse Mab to human TSH, than was seen with normal controls. However, F(ab')2 isolated from sera reacting with LE4 in the ELISA did not inhibit binding of the LE4 Mab with labeled TSH in a fluid phase competition assay. Patients with inactive Graves' disease showed higher ELISA reactivity with two different murine anti-TSH receptor Mabs than was recorded with either active Graves' or normal controls. A rough inverse correlation was noted between strongly positive ELISA reactions against these two Mabs with anti-TSH receptor specificity and the ability of immunoglobulins from inactive Graves' sera to stimulate increases in cyclic adenosine monophosphate (cAMP) in the normal rat thyroid cell line assay. Untreated Graves' sera showing high cAMP release only rarely showed elevated ELISA reactivity against Mabs with anti-TSH receptor activity.     

Methodological pitfalls in serum IgG2 level measurements by immunoenzymatic assays with monoclonal antibodies.

Four anti-IgG2 monoclonal antibodies (Mabs) were evaluated for their reactivity with purified myeloma IgG2 of different light chain types and Gm allotypes in three distinct immunoenzymatic assays (ELISA). The reactivity of three Mabs with solid-phase antigens was similar whereas an anti-Fab antibody (clone HP 6114) predominantly bound IgG2 kappa. In competitive and immunometric (sandwich type) assays, the binding of the two anti-IgG2 Mabs (HP 6014 and HP 6114) reacting with epitopes located on the Fab fragment was strongly influenced by the light chain type of IgG2 and by other factors (probably including differences in the variable regions); the Mab HP 6114 reacted virtually only with IgG2 kappa whereas the Mab HP 6014 displayed a much stronger affinity for IgG2 lambda than for IgG2 kappa; for both anti-Fab Mabs, important differences were found in their binding to individual IgG2 proteins. In addition, the Mab HP 6014 seemed to show a slightly better affinity for IgG2 bearing the Gm(23) allotype. These results urge much caution in IgG2 level measurement, especially with commercial kits, most of which use the Mab 6014 as the single anti-IgG2 reagent.     

Evaluation of the diagnostic value of measuring IgG,IgM, and IgA antibodies to mycobacterial A60 antigen in active tuberculosis

The purpose of the present study was to evaluate the clinical usefulness of detection of serum immunoglobulin A (IgA), IgG, and IgM antibodies raised against the mycobacterial A60 antigen for the diagnosis and discrimination of active tuberculosis (TB) from other pulmonary diseases. Three commercially available ELISA kits (IgA, IgG, and IgM) (ANDA Biologicals, Strasbourg, France) were evaluated simultaneously in 246 serum samples from 3 groups of patients: group I, 171 patients with active TB (128 pulmonary TB and 43 extrapulmonary TB); group II, 73 patients with pulmonary non-TB diseases; and group III, 2 leprosies patients. The sensitivities of tests ranged from 31.3% (IgA) to 94% (IgG) in pulmonary TB patients and from 21% (IgA) to 84% (IgG) in extrapulmonary TB patients. The specificities of assays varied from 92% (IgG) to 96% (IgA) in the pulmonary non-TB group. Combination of IgG with IgA and/or IgM does not improve its sensitivity. Clinical use of the A60-based serodiagnostic IgG assay is of great value for the rapid diagnosis and discrimination between active TB and pulmonary non-TB diseases. Moreover, this test could be used to increase diagnostic accuracy, especially for smear-negative TB cases, which are difficult to diagnose.     

Prevalence of human anti-mouse antibodies (HAMAs) in routine examinations

BackgroundCirculating heterophilic antibodies interfere with immunological assays in laboratory examinations; however, their rate of incidence is currently questionable. We developed an enzyme-linked immunosorbent assay (ELISA) to detect human anti-mouse antibodies (HAMAs) in routine examinations.MethodsThe study samples were comprised of serum samples obtained from 290 inpatients and outpatients at our hospital. Mouse immunoglobulin G1 (mIgG1), mIgG2a, and mIgG2b were used as the antigens and horseradish peroxidase (HRP)-conjugated anti-human IgG and IgM were used to identify the HAMA isotype.ResultsHAMAs were detected in 11.7% (34/290) of the samples. We observed 18 and 20 samples positive for IgG- and IgM-type HAMAs, respectively. Four samples contained both IgG- and IgM-type HAMAs. HAMAs against mIgG1, mIgG2a, and mIgG2b were found in 21, 14, and 13 samples, respectively. Existence of HAMAs was confirmed by western blotting using mIgG's as the antigens and HAMAs as the primary antibodies. Heterophilic blocking reagent (HBR) was also used to block the heterophilic interactions. Unexpectedly, a low HBR concentration rather enhanced the interactions instead of blocking them.ConclusionsA considerable number of HAMA-positive samples, reacting with the heavy chain of mIg, were found in routine examinations. A sufficient amount of HBR should be used for blocking the heterophilic interactions.     

Comparison of the effector functions of human immunoglobulins using a matched set of chimeric antibodies

Cell lines have been established that secrete a matched set of human chimeric IgM, IgG1, IgG2, IgG3, IgG4, IgE, and IgA2 antibodies that are directed against the hapten 4-hydroxy-3-nitrophenacetyl. These chimeric antibodies secreted from mouse plasmacytoma cells behave exactly like their authentic human counterparts in SDS-PAGE analysis, binding to protein A and in a wide range of serological assays. The antibodies have been compared in their ability to bind human C1q as well as in their efficacy in mediating lysis of human erythrocytes in the presence of human complement. A major conclusion to emerge is that whereas IgG3 bound C1q better than did IgG1, the chimeric IgG1 was much more effective than all the other IgG subclasses in complement-dependent hemolysis. The IgG1 antibody was also the most effective in mediating antibody-dependent cell-mediated cytotoxicity using both human effector and human target cells. These results suggest that IgG1 might be the favoured IgG subclass for therapeutic applications.     

Serum IgG, IgM and IgA concentration determined by the "linear plate" immunodiffusion technique in a normal population

Using the “linear plate” immunodiffusion technique, the authors have determined the serum IgG, IgM and IgA levels in normal populations. The incidence of storage of the samples at ?20° for a prolonged time was investigated and also the influence of age on immunoglobulin levels in a normal population. The reference sera were obtained by mixing a great number of serum samples (four pools were obtained for each immuniglobulin) containing the protein at markedly different levels. The immunoglobulin concentration of each pool was determined by comparison with a calibration curve obtained with well weighted quantities of immunochemical pure human IgG or IgM or IgA powder. The frequency distribution of the immunoglobulin levels in the populations investigated was found to be of the log normal type. The mean levels (1275 mg/100 ml for IgG; 74 mg/100 ml for IgM and 208 mg/100 ml for IgA) fit very well with the data reported by most authors using other methods. The determination of the IgG level was not influenced by storage of the samples at ?20°. A decrease of IgM and IgA was, however, observed when sera were kept frozen for a period of 3 months.No significant variation was found between the IgG and IgM levels of the different age classes. For the IgA's, however, the authors found a slight but significant increase between 20 and 30 years. Between the older age classes the increase is slower and between 40 and 60 there is no longer a significant change.     

Serodiagnosis of Lyme borreliosis using detection of different immunoglobulin (sub)classes by enzyme-linked immunosorbent assay and Western blotting

To improve the performance of enzyme-linked immunosorbent assays for the serodiagnosis of Lyme borreliosis, the prevalence of several immunoglobulin classes and subclasses against various antigens of Borrelia burgdorferi was investigated by Western blotting. The sera of 40 early Lyme borreliosis patients (ELB), 27 late Lyme borreliosis patients (LLB), 62 healthy controls and 140 non-Lyme borreliosis patients were used. Detection of IgG1 versus total IgG was found to be more sensitive in detecting Borrelia burgdorferi antigens, especially flagellin (41 kD) protein, but did not improve the performance of Western blotting. The use of IgG1 detection showed an increase in sensitivity and specificity for the early Lyme borreliosis patient group compared to the standard IgG and IgM detection method by enzyme immunoassays using purified Borrelia burgdorferi flagellum. However, in an enzyme immunoassay using a total sonicate, sensitivity in detecting early Lyme borreliosis and late Lyme borreliosis with IgG1 remained lower compared to the detection of early Lyme borreliosis by IgM antibodies and late Lyme borreliosis by total IgG antibodies.     

Heterophilic antibodies: a problem for all immunoassays

We verified that antibody-binding substances in serum that interfere in two-site immunoassays involving murine antibodies are heterophilic antibodies. Incubation of serum containing heterophilic antibodies and a murine monoclonal antibody to human choriogonadotropin (hCG) leads to formation of a series of soluble immune complexes. We investigated the recognition of hCG by reagent antibody in the presence of heterophilic antibodies and found this recognition to be diminished. Consequently, about 30% of serum samples containing heterophilic antibodies falsely appear to contain increased concentrations of hCG. The effect on analyte recognition probably results from steric inhibition of hCG binding to complexed antibody. Heterophilic antibodies detected with a murine antibody also bound immunoglobulin from several other species but did not bind all of those tested.     

Diagnostic value of anti-arc 5 IgG antibody and analysis of the IgG subclasses in sera of patients with cystic hydatid disease

Serodiagnosis of cystic hydatid disease (CHD) due to the metacestode of Echinococcus granulosus depends on detecting antibodies specific to hydatid antigen, but cross-reactivity with other parasites is one of the major draw-backs. We used a commercially-available antigen that elicits the arc 5 in immunoelectrophoresis and analysed the immunoglobulin class and the IgG subclass response by an ELISA. We tested sera from patients with confirmed CHD, cystic mass/lesions (CML) of non-hydatid origin, cysticercosis and healthy controls. High levels of antibodies to the hydatid antigen in all three classes (IgG, IgM and IgA) were observed only in the CHD patients. Significantly, only IgG antibody levels were discriminative and of diagnostic value. Also discussed is the validity of reading the significant cut-off point in IgG-ELISA in relation to a clinically important group of patients such as those with a CML of non-hydatid origin rather than healthy controls. Analysis of the anti-arc 5 IgG subclass responses demonstrated high antibody responses in all subclasses among the hydatid patients, with IgG3 the most discriminatory. The significance of the elevation of all four subclasses and more specifically of IgG4 in CHD is discussed in relation to certain biological activities of these immunoglobulin molecules.     

Profiles of brucella-specific immunoglobulin G subclasses in sera of patients with acute and chronic brucellosis

Serum specimens from patients with acute brucellosis (164), chronic brucellosis (22) and controls (75) were tested by ELISA for brucella-specific IgG, IgM, IgA and subclasses of IgG1 to 4 antibodies, Rose Bengal antigen slide agglutination (RB) and microagglutination (MA) tests. The RB and MA showed similar results and were positive in 100% and 64% of specimens from patients with acute and chronic brucellosis respectively. ELISA IgG and IgA were positive in sera from all patients with brucellosis while IgM was positive in 100% and 32% of specimens from patients with acute and chronic brucellosis respectively. Elevated IgG subclasses to brucella antigen were found in different proportions in the sera of patients with acute and chronic brucellosis. In patients with acute brucellosis, IgG1 was the predominant response (79%) followed by IgG3 (58%), IgG2 (36%) and IgG4 (14%). In contrast, IgG4 was the predominating subclass response (73%) in patients with chronic brucellosis followed by IgG1 (41%), IgG2 and IgG3 (27% each). When considering the most common elevated IgG subclasses, in each serum, either alone or in combination with each other, patients with acute brucellosis showed IgG1+IgG3 (24%), IgG1 (19%), IgG1+IgG2+IgG3 (16%) while patients with chronic brucellosis showed IgG4 (27%) and IgG1+IgG2+IgG3+IgG4 (18%). This study reveals that in addition to the difference in brucella-specific Ig class response in patients with acute (IgG, IgM, IgA) and chronic (IgG, IgA) brucellosis, the profiles of IgG subclasses are different where IgG1 predominates in the acute and IgG4 in the chronic stages of the disease.     

Beneficial role of endogenous immunoglobulin subclasses and isotypes in septic shock

Purpose

There is increasing evidence on the relationship between endogenously produced immunoglobulins and the clinical outcome in septic shock (SS).

Materials and methods

Levels of immunoglobulin G (IgG) subclasses, immunoglobulin A (IgA), immunoglobulin M (IgM), and immunoglobulin E were measured in plasma from 42 patients with SS and in 36 patients with systemic inflammatory response syndrome at diagnosis. Association of immunoglobulins levels with disease severity and outcome was evaluated.

Results

Eighteen patients with SS finally died. Both patients with systemic inflammatory response syndrome and SS showed subnormal levels of total IgG, IgG2, and IgM. Patients with SS who died showed the lowest levels of total IgG and IgG1. Total IgG, IgG1, IgG2, IgG3, IgG4, and IgA correlated inversely with Acute Physiology and Chronic Health Evaluation II score in SS. Univariate Cox regression analysis showed that levels of IgG1, IgG2, IgG3, IgM, IgA, and total IgG were inversely associated to the probability of death at 28 days. Multivariate analysis showed that IgG1, total IgG, IgM, and IgA behaved as independent protective factors against mortality (hazard ratio, P): 0.23, 0.026; 0.16, 0.028; 0.11, 0.042; 0.05, 0.010, respectively, whereas IgG3 showed a protective trend also.

Conclusions

Our study evidenced that, in addition to IgG1, other major endogenous immunoglobulins isotypes and subclasses seem to play a beneficial role in SS.     

Enzyme-linked immunosorbent assay for determination of antibodies to xanthine oxidase

An enzyme-linked immunosorbent assay (ELISA) was developed to detect IgG, IgA and IgM antibodies to xanthine oxidase. The method used xanthine oxidase to coat sample wells on microtitre plates. The anti-xanthine oxidase concentrations were determined by reference to standard curves constructed by coating plates with anti-IgG, anti-IgA and anti-IgM to capture antibodies of different classes in standard human serum. The standard curves for IgG, IgA and IgM had a working range of 0 to about 60 ng/ml, and all results with commercial quality control serum fell within expected ranges. The coefficients of variation (CV) for within-batch precision (n = 30) and between-batch precision (n = 20) for IgG and IgM were about 9% and 12% respectively. The detection limit was 2 ng/ml. The ELISA was applied to assay serum samples of 110 Chinese and 110 European healthy subjects. A positively-skewed distribution in their anti-xanthine oxidase IgG and IgM levels was observed.     

Human anti-animal antibody interferences in immunological assays.

Purpose: The scope and significance of human anti-animal antibody interference in immunological assays is reviewed with an emphasis on human anti-animal immunoglobulins, particularly human anti-mouse antibodies (HAMAs). Issues: Anti-animal antibodies (IgG, IgA, IgM, IgE class, anti-isotype, and anti-idiotype specificity) arise as a result of iatrogenic and noniatrogenic causes and include human anti-mouse, -rabbit, -goat, -sheep, -cow, -pig, -rat, and -horse antibodies and antibodies with mixed specificity. Circulating antibodies can reach gram per liter concentrations and may persist for years. Prevalence estimates for anti-animal antibodies in the general population vary widely and range from <1% to 80%. Human anti-animal antibodies cause interferences in immunological assays. The most common human anti-animal antibody interferent is HAMA, which causes both positive and negative interferences in two-site mouse monoclonal antibody-based assays. Strategies to prevent the development of human anti-animal antibody responses include immunosuppressant therapy and the use of humanized, polyethylene glycolylated, or Fab fragments of antibody agents. Sample pretreatment or assay redesign can eliminate immunoassay interferences caused by anti-animal antibodies. Enzyme immunoassays, immunoradiometric assays, immunofluorescence, and HPLC assays have been designed to detect HAMA and other anti-animal antibodies, but intermethod comparability is complicated by differences in assay specificity and lack of standardization. Conclusions: Human anti-animal antibodies often go unnoticed, to the detriment of patient care. A heightened awareness on the part of laboratory staff and clinicians of the problems caused by this type of interference in routine immunoassay tests is desirable. Efforts should be directed at improving methods for identifying and eliminating this type of analytical interference.     

抗核抗体荧光模式及滴度与免疫球蛋白及补体的关系分析

目的探讨抗核抗体荧光模式及滴度与免疫球蛋白及补体的关系。方法采用透射比浊法对768例抗核抗体阳性患者血清进行免疫球蛋白及补体检测,另选取200名健康体检者作为对照组。根据抗核抗体荧光模式及滴度进行分组,分别研究各组免疫球蛋白及补体水平。结果①抗核抗体滴度≥1∶100时,血清IgG、IgA明显升高(P<0.01),但升高的幅度与抗核抗体滴度无明显相关;IgM和C3、C4只有在抗核抗体滴度≥1∶1 000时才分别显著升高和下降(P<0.01)。②各种单荧光模式型患者血清IgG、IgA、IgM、C3、C4浓度无显著差异;与各单荧光模式患者比较,二联及二联以上荧光模式型患者血清IgG、IgA、IgM显著升高、补体C3、C4水平显著下降(P<0.01)。结论抗核抗体、免疫球蛋白及补体虽然与自身免疫性疾病高度相关,但并不与病情成平行关系,三项联合检测才能对疾病的诊断与病情发展作出正确判断。     

Immunoglobulins and IgG subclasses in children following severe head injury

Objective To study immunoglobulin production after severe blunt head trauma in children.Design Serum for IgG, IgM, IgA, IgE, and IgG subclasses were drawn from 10 children admitted with severe head injury (ISS 31.2, GCS 5.4) on day 1, 7, 14 and 21 after injury.Results 5 of the 10 patients developed infection between 7 and 14 days and 2 died of complications of pneumonia. On day 1, IgM levels averaged 95.6% of the mean of the age-specific normal controls. By day 7, IgM levels averaged 383% (p<0.01). While all patients were within the age-specific normal range (± 2 SD) on day 1, 7 of 10 patients were above the normal range by day 7. There was no difference in IgM levels between infected and non-infected patients. Five patients were below the age-specific normal range for IgG on day 1, with 3 still low on day 7. By day 21, IgG levels averaged 141% of the mean of the age-specific normal controls. IgG subclasses followed a pattern similar to total IgG levels. Marked increases in IgE were seen in 3 patients.Conclusions IgM levels increased dramatically in all patients within seven days of the injury. While 50% of these children had a deficit of IgG in the first week, total IgG and IgA levels increased after injury, but not as rapidly as IgM levels. Unlike pediatric burn patients, there is no persistent hypogammaglobulinemia following severe blunt trauma in children.Supported by a grant from Sandoz Pharmaceuticals, East Hanover, New Jersey, USA     

Antibodies against acetaldehyde-modified epitopes: an elevated IgA response in alcoholics

Several recent reports have shown that antibodies reactive with acetaldehyde (AcH)-modified epitopes are present in alcoholics. However, similar antibodies have also been found in patients with non-alcoholic liver disease and control subjects. In each of these studies total immunoglobulin binding to the AcH-modified proteins was measured, with no attempt being made to identify the classes of immunoglobulin involved. In the present study we employed an enzyme-linked immunosorbent assay (ELISA) to assess the classes of immunoglobulin involved in this response, using plasma samples from 97 alcoholics with varying degrees of liver disease, 35 patients with non-alcoholic liver disease and 33 control subjects. All three groups exhibited a large IgM response and a negligible IgG response. However, the alcoholics exhibited a significantly higher IgA response than either of the other groups. This suggests that the measurement of the IgA response to AcH-modified epitopes may be a specific marker of ethanol abuse.