HBsAg as the antigen component of circulating immune complexes in HIV-infected patients.

Seeing the same transmission pattern of HIV and HBV coinfection by these two agents is not an uncommon feature. Immunity impairment due to HIV infection can be the cause of a higher rate of HBV replication with less intensive liver damage and less effective immune response to HBV, while the pathological course in both infections involves elevated levels of circulating immune complexes (CIC). These were the reasons for us to examine the frequency of HBsAg involvement as the antigen component of circulating immune complexes formed in sera of HIV-infected patients in different stages of HIV disease. We tested 67 sera of HIV-positive patients in different stages of HIV disease for the presence of HBsAg and HIV antigen p24 (with and without acid dissociation of immune complexes), for the presence of anti-Hbc antibodies and circulating immune complexes. HBsAg was positive in 13.8% sera prior to and 33.8% after acid pretreatment. Anti-HBc antibodies were present in 76.9% serum samples tested. Fifty percent of sera were positive for both HBsAg and p24 antigen after dissociation of immune complexes. The level of CIC was elevated in 65.9% of sera. Our results suggest that HBsAg is commonly associated in immune complexes formed in the sera of HIV-infected patients and that they may simultaneously contain HIV and HBsAg in patients coinfected with both agents. This may contribute to their mutual interaction and influence the diagnosis and follow-up of patients.     

Characterization of autoantibody activities in sera anti-DNA antibody and circulating immune complexes from 12 systemic lupus erythematosus patients

To examine autoantibodies present in patients with active systemic lupus erythematosus (SLE), sera, circulating immune complexes (CIC), and antibodies purified on DNA-immunoadsorbent were tested by enzyme immunoassay. A panel of self-antigens, including DNA, histones (HIS), glomerular basal membrane (GBM), thymus cell extract (TCE), actin (ACT), myosin (MS), and tubulin (TUB), was used to define their specificities. IgM antibodies against all antigens of the panel were detected in sera, CIC, and in antibodies eluted from the DNA-immunoadsorbent and demonstrated a large polyreactivity. IgG antibodies showed restricted activities against DNA, HIS, GBM, and TCE in sera and a large polyreactivity in CIC. Inhibition experiments were performed to assess their mono- or polyreactivities. Among the IgG autoantibody population recognizing DNA, two populations of IgG antibodies were detected in the sera and in the affinity purified anti-DNA: one recognizes DNA, HIS, and GBM, and the other binds to DNA and to cytoskeletal proteins. These autoantibody populations were found in CIC, which also often contained high amounts of IgG antibodies recognizing ACT and MS. A third population of IgG antibody that recognizes only TCE and could not be inhibited by DNA or other antigens was found in serum and CIC. Our data demonstrate the existence of several populations of autoantibody in serum and CIC of SLE patients: (1) IgM polyreactive autoantibodies, (2) IgG polyreactive autoantibodies recognizing DNA and cytoskeletal proteins, (3) IgG specific to DNA, which cross react with HIS and GBM, and (4) IgG specific to TCE antigens. © 1996 Wiley-Liss, Inc.     

Disorders of humoral immunity in hypertension

Examination included 51 patients with essential hypertension (EH) of an uncomplicated course, labile (IB-IIA stage, according to A. L. Myasnikov's classification) and stable (IIB stage) hypertension. Clinical characteristics were given to the stages and duration of EH, body weight of the patients, arterial hypertension (AH) heredity and the AP level. Immunological examination included determination of the concentration of the basic classes of immunoglobulins IgG, IgA, IgM, circulating immune complexes (CIC), concentration of IgE and beta 2-microglobulins. It was revealed that EH development is attended by an increased concentration of immunoglobulins, primarily of IgA (23 per cent), IgE (31 per cent) and CIC (21 per cent), which is associated, to a certain degree, with a factor of AH hereditary aggravation.     

Erroneous results with routine laboratory testing for immunoglobulins due to interference from circulating immune complexes in a case of hyperviscosity syndrome associated with autoimmune disease

A patient with the hyperviscosity syndrome exhibited very high concentrations of intermediate to small circulating immune complexes (CIC), involving 40-50% of the IgG present, with IgG rheumatoid factor activity. We demonstrate that precipitation of CIC by polyethylene glycol in the reaction mixture caused interference with nephelometric methods for measuring IgM and IgA, and that failure of immunoglobulins to migrate, owing to molecular interactions, caused interference with radial immunodiffusion methods. Semiquantitative values for immunoglobulins were difficult to interpret on immunoelectrophoresis. As a result, IgM and IgA could only be quantitatively estimated by an end-point nephelometric approach that included a serum blank. Immunoelectrophoresis indicated that a large proportion of the immunoglobulins behaved as aggregates. Immunofixation electrophoresis did not reveal the presence of aggregates. The polyethylene glycol-IgG test provided an accurate assessment of the CIC concentration; the Raji cell and C1q-binding assays did not. Evidently, special techniques may be necessary for accurate determination of immunoglobulin concentrations when CIC concentrations are very high.     

Circulating immune complexes in pulmonary tuberculosis

The presence of circulating Immune complexes (CIC) in sera from patients with pulmonary tuberculosis was demonstrated by 3 techniques (a) latex agglutination (b) 3.5% PEG precipitation and determination of optical density (O.D.) at 280 nm and (c) radioimmunoassay (RIA) of CIC using bovine spermatozoa. 40 normal control sera and 100 T.B. patients sera were included in the study. 12% cases were positive for CIC by all the 3 methods mentioned above, 13% were negative by all the 3 methods and the remaining patients were positive by one or more methods of detection. To correlate the levels of CIC as detected by different techniques with the activity of the disease, patients were broadly grouped as (a) sputum positive and (b) sputum negative. Higher levels of CIC were obtained in sputum positive cases than sputum negative by all the 3 methods studied. While IgG, IgA and IgM were elevated in the CIC of T.B. patients, and IgG and IgA were also present in controls, IgM immunoglobulins were detected only in patients and not in controls. The effect of antitubercular treatment on the levels of CIC was also evaluated and it was found that the levels of CIC remained unchanged even after prolonged chemotherapy.     

Immunological features of renal lesion in chronic alcoholism

One hundred and sixty males whose mean age was 42 years were examined. Of them there were 122 patients with stage II chronic alcoholism (CA), 92 with renal lesion following the type of chronic glomerulonephritis (CG) (Group 1); 30 patients with CA without renal lesion (Group 2), and 42 patients had CG alone (Group 3). Methods that characterize humoral immunity were used. These included detection of circulating immune complexes (CIC) by polyethylene glycol precipitation, measurement of the concentrations of IgA, IgM, and IgG by the Mancini radial immunodiffusion assay, detection DNA antibodies by the Farre test modified by V. V. Koshelev, that of serum anticomplement activity, measurement of the levels of complement by its hemolytic activity, determination of the activity of the lysosomal enzymes acid RNAase, acid DNAase, and cathepsin by the procedure of A. A. Pokrovsky et al. Complex estimation of the content of CIC, immunoglobulins, DNA antibodies and the activity of the lysosomal enzymes in patients with renal lesion makes it possible to evaluate the severity of a pathological process and to make its prognosis.     

Circulating immune complex levels measured by new ELISA kits utilizing monoclonal anti-C1q and anti-C3d antibodies correlate with clinical activities of SLE but not with those of RA

The authors studied sera from both patients with SLE and from those with RA to evaluate clinical usefulness and significance of circulating immune complexes (CIC) detected with new ELISA kits utilizing monoclonal anti-C1q and anti-C3d antibodies. CIC values of patients with SLE significantly correlated to severity of disease activities evaluated by clinical symptoms and laboratory tests, especially for serum complement levels. Since substances detected with the ELISA kits were closely related to serum complement components, it was determined that a direct relationship exists between clinical activities and CIC values appearing in SLE patients with hypocomplementemia. In RA patients, CIC values did not correlate to clinical activities evaluated by Lansbury's index, anatomical bone damage with X-ray or functional assessment of activities of daily living, but did significantly correlate to levels of IgM-RF, serum IgG concentrations and some markers of systemic inflammation. Detection of IC after fractionation of RA sera revealed a broad range of molecular sizes detectable with the ELISA kits, which indicated that CIC in vivo were heterogenic and complicated in formation, degradation and interaction with serum complements.     

Antibodies to the myelin major protein in patients with dilated cardiomyopathy. Determination and diagnostic significance

IgG and IgM antibodies to the myelin major protein (MMP) have been measured by solid-phase enzyme immunoassay (EIA) in the blood sera of 31 patients with dilatation cardiomyopathy (DCMP), 4 ones with coronary disease (CD), and 20 normal subjects. Nonspecific binding of the native and aggregated human gamma-globulin with the MMP has been examined, as well as the correlation between the blood serum concentrations of circulating immune complexes, IgG and IgM, and the rheumatoid factor, and the titers of IgG and IgM antibodies to MMP; no correlation has been detected (p greater than 0.05). The activities of bovine and porcine MMP in EIA have been under study and found identical (p less than 0.001). No anti-MMP antibody has been detected in CD patients, whereas in DCMP IgG antibodies have been present in 55% and IgM ones in 49% of the cases, and 52% of the patients had both types of antibodies. Dynamic titration of anti-MMP antibody carried out in 9 DCMP patients has revealed a 4-fold and higher elevation of the titers in 4 patients and a reduction of these titers in other 4 patients. Comparison of the titers of IgG and IgM antibodies to MMP and the peripheral blood lymphocyte ability to spontaneous synthesis of immunoglobulins and Phytolacca mitogen-stimulated Ig synthesis by the MMP has revealed a correlation between these parameters. The findings permit a hypothesis on the involvement of the nerve fiber myelin-containing cells, possibly of the afferent system of the heart, in half the examinees with DCMP.     

Immunoglobulin isotypes in the circulating immune complexes of diabetic rats with and without proteinuria: a chronological study

Circulating immune complexes (CIC) have been postulated to contribute to the development of the secondary complications of diabetes mellitus. Therefore studies were performed to determine whether CIC are the cause of the consequence of the development of diabetic nephropathy. This was done by comparing the occurrence and concentration of CIC containing the different isotypes of immunoglobulins in control rats to those detected in streptozotocin-induced (Sz) diabetic rats that developed albuminuria (Group I) and that did not develop albuminuria (Group II). Only CIC containing IgM, IgG2b and IgG2c were detected in diabetic rats. By staging Group I albuminuric diabetic rats to a clinical reference point of albuminuria, there was no correlation between the occurrence or concentration of CIC containing any isotype of immunoglobulin and the onset of albuminuria. In all Group I albuminuric diabetic rats, the occurrences of all CIC were variable and their concentrations fluctuated during the development and early progression of nephropathy. However, after this group of diabetic rats progressed to overt nephropathy (marked by albuminuria and IgG proteinuria), CIC could be demonstrated in 100% of the animals. In diabetic rats that did not develop albuminuria (Group II), CIC containing IgG2b occurred earlier and more often than in Group I albuminuric rats. Similarly, the subclass IgG2c was detected in the CIC of Group II non-albuminuric rats more frequently and in higher concentrations than in Group I albuminuric rats. CIC containing IgM were detected in all 3 animal groups, however, in higher concentrations and occurrences in Group II non-albuminuric rats as compared to control and Group I albuminuric rats. The consistent elevation in CIC after the development of diabetic nephropathy, suggests that the CIC containing any immunoglobulin isotype either result from diabetic kidney, or from other deteriorating conditions associated with the diabetic state. The data also suggests that CIC are not involved in the onset or progression of diabetic nephropathy regardless of the isotypes of immunoglobulins contained within the CIC. However, there is an isotypic restriction in the immunoglobulins detected in the CIC of diabetic rats (IgM, IgG2b and IgG2c) that may signal some involvement of the immune system in the development of diabetic nephropathy.     

Circulating immune complexes in sarcoidosis

Circulating immune complexes (CIC) were detected in 100 out of 112 sera from 33 sarcoidosis patients. Five tests were used representing three different basic principles. All patients had detectable CIC at some stage of their disease. The three platelet tests detecting IgG complexes exhibited the highest positive titres in the acute cases. The ClqB-ELISA test, which detects complement fixing IgG complexes, was the test most frequently positive in the chronic cases. The presence of extra-pulmonary lesions or corticosteroid therapy did not influence the appearance or disappearance of CIC. No positive correlation was found between CIC and elevated levels of serum angiotensin-converting enzyme (ACE) and/or serum lysozyme (LZM). The evaluation of CIC in sarcoidosis requires a battery of different tests carried out at regular intervals during the follow up.     

Immunoglobulins from the sera of immunologically activated persons with pairs of electrophoretic restricted bands show a greater tendency to aggregate than normal

From in vitro data, it has been speculated that pairs of endogenous restricted bands migrating in close proximity in the gamma region upon high resolution serum electrophoresis (HRE) represent circulating immune complexes (CIC). Using a polyethylene glycol (PEG) method to separate CIC, we found a very high correlation between the presence of such band pairs and elevated levels of CIC (CHI2 = 25.7, p less than 0.001) in 51 sera. HRE appears to be a good screening technique to identify, with a high degree of certainty, samples with elevated levels of CIC for delineation by more specific methods. Yet, examination by gel-filtration chromatography and precipitation with PEG indicated that the molecules comprising the band patterns were not CIC, but polyclonal 7S IgG. These bands are usually found in patients with chronic activation of the immune system. Fractionation of the sera from 5 such patients with various cuts of PEG indicated that the average IgG concentration in the 2.5-5%, and 5-7.5% cuts from patients was 3.99 g/l and 2.2 g/l, while from healthy subjects the concentrations were 0.68 g/l and 2.88 g/l. This reversed precipitation pattern was seen both for absolute levels of IgG and for percent of total IgG. On the average the amount of precipitation of IgG in the 2.5-5% fraction of patients was about 5-fold above that seen in the healthy subjects. The endogenous bands were not associated with any specific cut of PEG, but appeared to be proportionally distributed in accord with the levels of IgG. The data is consistent with the idea that immunologically activated patients exhibit a greater tendency for immunoglobulins to associate than normal. This propensity to aggregate may cause CIC to form in situ in local compartments even though CIC do not appear to be present upon analysis by biochemical techniques.     

Antigen-specific immunoglobulins in patients with acute pulmonary, chronic pulmonary, or disseminated histoplasmosis

Patients infected with Histoplasma capsulatum exhibit protean clinical manifestations and similarly express variable humoral immune responses. Therefore, the specific goals of this study were to more clearly define host immune responses by determining the concentrations of total and H. capsulatum-specific immunoglobulins in sera from patients with acute or chronic pulmonary and disseminated histoplasmosis. H. capsulatum-specific (AS) IgG, IgA and IgM, and total IgE were determined by radioimmunoassays while total IgG, IgA, and IgM were quantitated by laser nephelometry. In general, total IgG and IgA were elevated, while IgM and IgE were either decreased or normal for the three clinical forms of histoplasmosis studies. Antigen-specific IgG and IgA were markedly elevated in all classes of disease, whereas AS-IgM was only slightly increased. Total and AS-IgG were elevated in the sera of chronic patients directly proportional to the number of demonstrable precipitin bands.     

Anti-Fab antibodies in humans. Predominance of minor immunoglobulin G subclasses in rheumatoid arthritis

Isoelectric focusing analyses of sera from patients with rheumatoid arthritis (RA) demonstrate two populations of antibodies directed against the Fab portion of pooled human IgG. One population is composed of polyclonal alkaline anti-Fab antibodies (alpha FABA) and the other, acidic alpha FABA which are more clonally restricted. In this study we have identified the immunoglobulin classes and subclasses of these antibodies in RA sera. Enzyme-linked immunosorbent assays (ELISA) demonstrated alpha FABA in RA sera to be predominantly IgG. A large portion of IgG alpha FABA existed as immune complexes, inasmuch as dialysis of RA sera against 6 M urea before ELISA analysis was necessary for maximal detection of alpha FABA activity. Chromatofocusing of RA sera isolated alpha FABA of different charges and revealed the acidic clonally restricted alpha FABA to be IgG4 and IgG3, whereas the polyclonal alkaline group contained IgG1, IgG2, and IgG3. Overall, acidic IgG3 and IgG4 comprised 70% of IgG alpha FABA, and high levels of IgG4 were seen in most RA sera. When alpha FABA were elevated in normal sera, they were primarily of the IgG4 subclass, and also existed as immune complexes. Serum anti-Fab activity was removed by adsorption of sera with Fab fragments. Anti-Fab antibodies of both kappa and lambda light-chain types were present in RA sera, and F(ab')2 fragments of RA serum immunoglobulin were found to possess anti-Fab activity. These studies indicate that alpha FABA in RA sera are limited to the IgG class, and that most of these antibodies exist as immune complexes and display clonal and minor IgG subclass restriction.     

The nature and incidence of cryoproteins in hepatitis B antigen (HbsAg) positive patients.

Hepatitis B (HbsAg) surface antigen has been detected in the serum of patients with a variety of diseases and immune complexes of this antigen and antibody have been implicated in tissue damage to various organs. Previously we have demonstrated that serum cryoproteins occur in a variety of immune complex disorders and represent pathogenic complexes of antigen and specific antibody. Sera from patients with acute HbsAg positive hepatitis, chronic hepatitis B antigenemia, acute and chronic HbsAg negative hepatitis, as well as a variety HbsAg negative miscellaneous liver diseases and normals were studied for the presence and nature of cryoproteins. Cryoproteins were detected in a large number of patients with acute and chronic HbsAg positive hepatitis and chronic HbsAg carriers. The quantity of these cold insoluble precipitates was highest in acute hepatitis. Cryoproteins were detected with much less frequency in HbsAg negative patients and were not found in normals. The precipitates in HbsAg patients contained either HbsAg, anti-HBsAg or both, along with immunoglobulins and occasionally complement and rheumatoid factor. The cryoproteins in these patients had biological properties attributable to immune complexes and several of the patients had clinical manifestations of acute or chronic serum sickness. Cryoproteins from HbsAg negative patients did not contain HbsAg or antibody to HbsAg and did not have biologic properties of immune complexes. In HbsAg positive patients HbsAg and antibody to HbsAg were concentrated in the cryoprecipitate. The preliminary studies suggest that investigation on cryoproteins in hepatitis may be of clinical and immunopathogenic value.     

The detection of circulating immune complexes containing immunoglobulin G

An immune complex assay using radiolabelled immunospecific antibodies against human IgG and polyethylene glycol precipitation (IgG-PEG assay) is described. The material reactive in this assay was evaluated using aggregated immunoglobulins, immune complexes prepared in vitro, sera of patients with a variety of disorders and normal human serum. Sucrose density gradient ultracentrifugation showed that only large-sized immune complexes (greater than 25 S) were detected. Comparison of the results of the IgG-PEG assay with those of the C1q binding assay showed a highly significant positive correlation (p less than 0.001). It was found that rheumatoid factors do not influence the results of the IgG-PEG assay. The method described in this study detects specifically immune complexes containing IgG and might be extended to the detection of other constituents of circulating immune complexes.     

Antibiodies of the IgG, IgM, and IgA classes in newborn and adult sera reactive with gram-negative bacteria

Umbilical cord serum and adult serum antibodies reactive with heat-stable somatic antigens of Gram-negative bacteria (Neisseria gonorrhoeae, Escherichia coli, and Salmonella typhosa) were assayed by using an indirect fluorescent antibody test. Reactive IgG, IgM, and IgA antibodies were identified by using fluoresceinconjugated antisera specific for these immunoglobulin classes.IgG antibody titers in cord serum approximated those found in the corresponding maternal sera. IgM and IgA antibodies were present in adult sera but were not demonstrable or were present only in small amounts in cord sera. The presence of IgG and IgM antibodies reactive with Gram-negative bacteria was confirmed by the testing of purified 7S and 19S fractions. In addition, both IgG and IgM reactivities were inhibited by the prior incubation of serum with purified specific lipopolysaccharide preparations.The ubiquity and magnitude of these natural IgG antibodies in the sera of both adults and neonates have apparently eluded detection in previous studies. The use of bactericidal and agglutination tests, which are apparently more sensitive to the presence of IgM than to IgG antibodies, may account for the failure of previous studies to detect adult and cord IgG antibodies reactive with somatic antigens of Gram-negative bacteria. The presence of these IgG antibodies may be correlated with the resistance to infection demonstrated by most newborns as they are challenged by the septic extrauterine environment.     

Autoantibodies in the sera of Trypanosoma cruzi-infected individuals with or without clinical Chagas disease

Variations in the levels and the specificities of autoantibodies directed against a panel of antigens (cytoskeleton proteins, DNA, laminin) were analyzed in the sera from two groups of humans infected with Trypanosoma cruzi. One group was constituted of apparently healthy blood donors (BD) and the other of patients with clinically confirmed Chagas disease (CCH). In both infected groups, a high proportion but not all sera exhibited dramatic enhancement of IgM and IgG autoantibodies directed against all antigens tested. Sera positive for IgG autoantibodies were generally found more frequently in the CCH than in the BD group, except for anti-actin antibodies more often present in BD sera. Anti-laminin IgG antibodies were present in a similar number of individuals in both groups. Although the titers of anti-laminin IgG antibodies were in general higher in CCH, their dissociation constants were in the same range (7 × 10^?8–10^?7M) in both groups. IgG autoantibodies were demonstrated to be polyreactive with laminin and other self antigens as well. Circulating immune complexes were present in sera from both groups and the activity of the antibodies dissociated from these complexes was directed against all the antigens of the panel. Although the IgE concentration was significantly enhanced in several subjects from both groups, the incidence of positive sera was higher in the CCH (60%) than in the BD (39%) group. Our results demonstrate that autoantibodies with the characteristics of natural autoantibodies are found in both T. cruzi-infected apparently healthy individuals and patients. © 1993 Wiley-Liss, Inc.     

Immune mechanisms in heparin-induced thrombocytopenia: no evidence for immunoglobulin M anti-idiotype antibodies

BACKGROUND: Heparin-induced thrombocytopenia (HIT), which is caused by platelet (PLT)-activating immunoglobulin (Ig)G antibodies against platelet factor 4 (PF4)/heparin complexes, differs from other immune responses seen in immunohematology: IgG antibodies are formed as early as 5 days even without previous heparin exposure; antibodies are remarkably transient (<100 days); HIT is more frequent in postsurgery patients compared with medical patients despite administering the same type and dose of heparin; and increasing evidence implicates autoantibody-like reactivity of anti-PF4/heparin antibodies. We hypothesized that these unusual features could be caused by loss of regulatory anti-idiotype IgM antibodies due to disturbance (e.g., by surgery) of an idiotype–anti-idiotype network.
STUDY DESIGN AND METHODS: Sera were obtained prospectively before heparin administration and during the immunization phase of HIT and also from patients with previous HIT after waning of antibodies to nondetectable levels. To detect inhibitory IgM anti-idiotype antibodies, we performed serum coincubation experiments and IgG purification by protein G and size filtration to exclude coprecipitating IgM. Sera (n = 3) containing known anti-PF4/heparin IgG or IgM antibodies and normal sera (n = 20) were processed as controls.
RESULTS: Fifteen preimmune response sera (seroconverting in the PF4/heparin-IgG enzyme-linked immunosorbent assay only [n = 4] or additionally in a PLT activation assay [n = 5] or in both assays plus thrombosis [n = 6]) and four sera of previously immunized patients were included. Neither did the neat sera inhibit binding of anti-PF4/heparin antibodies nor did the purified IgG fractions show enhanced binding to PF4/heparin complexes.
CONCLUSION: The atypical immunologic features of HIT do not appear to be caused by disruption of an idiotype (IgG)–anti-idiotype (IgM) network.     

Immunological characteristics of patients with acute coronary syndrome (unstable angina and myocardial infarction)]

Admission and discharge values of rosette formation, adhesive and phagocytic ability of neutrophils, exercise tests with calculation of the tension index (TI), serum concentrations of IgG, IgA and IgM, circulating immune complexes (CIC) and cardiolipin antibodies (CAB) were studied in 48 males with unstable angina pectoris and acute myocardial infarction. Acute coronary syndrome is shown to be associated with marked immune alterations, primarily, with elevated levels of CIC and CAB, reduced TI. These alterations persisted for 3-5 weeks of the hospital stay and provoked the risk of repeated infarctions and thrombotic complications after relief of clinical symptoms of acute coronary failure.     

Binding of complement component C5 to model immune complexes and the use of anti-C5 antibodies for determination of C5-containing circulating immune complexes

When normal human or mouse serum is added to micro ELISA plates coated with monomeric or aggregated IgG, complement component C5 binds to IgG. C5 binding was demonstrated with a specific chicken anti-C5 antibody. Hydrazine treatment of the serum or addition of EDTA to the serum abolished the binding of C5. C5-deficient mouse serum was negative for C5 binding, whereas the same serum supplemented with human C5 restored the binding of C5. Chicken anti-C5-coated plates were used for determination of C5-containing circulating immune complexes (CIC). Increased concentrations of CIC were found in sera from patients with rheumatoid arthritis and Bell's palsy.