Abstract
Background
Recent data suggest intestinal immunity including immunoglobulin A (IgA) may contribute to the pathogenesis of alcohol-induced liver disease (ALD).
Methods
We documented serum IgA levels in ALD patients and determined whether those with elevated levels of IgA (E-IgA) had similar, more, or less advanced disease and different rates of progression than those with normal levels of IgA (N-IgA). Standard liver function tests (bilirubin, international normalized ratio [INR], and albumin), model for end-stage liver disease (MELD), and Fibrosis-4 (FIB-4) scores were used as indicators of disease severity.
Results
From the study centre’s clinical database, we identified 175 adult patients with ALD, 107 (61%) with E-IgA and 68 (39%) with N-IgA. Gender distribution and mean age of the two cohorts were similar. E-IgA patients had biochemical evidence of more advanced liver disease (higher serum bilirubin and INR and lower albumin levels) than N-IgA patients (ps < .05). E-IgA patients also had significantly higher median MELD and FIB-4 scores (ps < .01). A higher percentage of E-IgA patients had FIB-4 values in keeping with advanced fibrosis or cirrhosis (55% versus 28%, p = .02). After mean follow-up periods of approximately 4 years, liver biochemistry and MELD and FIB-4 scores changed to similar extents in the two cohorts.
Conclusions
Serum IgA levels were increased in approximately 70% of ALD patients. Although these patients had biochemical and non-invasive indicators of more advanced disease, elevations in serum IgA levels do not predict disease progression; therefore, IgA is unlikely to be of importance in the pathogenesis of ALD.
Keywords: alcohol, alcohol-induced liver disease, alcoholic cirrhosis, alcoholic hepatitis, IgA, immunoglobulins
Introduction
Approximately two-thirds of US adults drink alcohol, and 5% meet established criteria for alcohol abuse or dependence (1). Given the high prevalence of alcohol consumption, it is not surprising that alcohol-induced liver disease (ALD) is a major contributor to disease morbidity and mortality. Indeed, one-half of deaths from cirrhosis, the 12th leading cause of death in the United States, can be attributed to excess alcohol consumption (2).
Several hypotheses have been advanced to explain the pathogenesis of ALD, including direct alcohol toxicity, ischemic injury, and immune-mediated hepatitis (3). Recently, attention has focused on alcohol-induced disruption of intestinal mucosal integrity, enhanced bowel permeability, increased lipopolysaccharide entry into the portal circulation, and activation of proinflammatory cytokine expression by fixed tissue macrophages in the liver (4–9). The latter hypothesis has resulted in a reassessment of the role that the enteric immune system plays in the pathogenesis of ALD.
Immunoglobulin A (IgA) is a subclass of immunoglobulins produced by B lymphocytes in various regions of the body, including the Peyer’s patches of the intestinal mucosa (10). Previous studies have documented high serum IgA levels in individuals who drink alcohol to excess and a heavy deposition of IgA in the sinusoids of alcohol-damaged livers (11–13). These findings, together with the ability of IgA to activate complement and initiate inflammatory reactions, raises the possibility that IgA contributes to the pathogenesis of ALD (14). Conversely, IgA derived from the intestinal mucosa possesses anti-lipopolysaccharide properties and therefore could play a role in attenuating the extent of liver injury in patients with ALD (15).
The objectives of this study were to document the prevalence of elevated serum IgA levels in patients with ALD and to determine whether associations exist between elevated serum IgA (E-IgA) levels and the severity of liver disease and whether baseline serum IgA levels predict the subsequent course of ALD.
Methods
Study database
The Section of Hepatology at the University of Manitoba maintains the Phil and Ellie Kives Clinical Database, a computerized patient database that captures demographic, clinical, laboratory, radiologic, and histologic data from referred adults attending the Liver Diseases Outpatient Clinics at the Health Sciences Centre, a tertiary health care centre in Winnipeg, Manitoba.
Definitions
For the purpose of this study, a diagnosis of ALD was established on the basis of a self-reported history of excess alcohol consumption (>40 g/day or 280 g/wk for men and >30 g/day or 210 g/wk for women), a serum aspartate aminotransferase/alanine aminotransferase (AST/ALT) ratio greater than 1, and gamma-glutamyl transferase (GGT) levels disproportionately elevated relative to alkaline phosphatase (ALP) levels in the absence of medications known to elevate GGT values.
Disease severity was defined by serum bilirubin, albumin, and international normalized ratio (INR) values, mean model for end-stage liver disease (MELD) scores, and median Fibrosis-4 (FIB-4) scores. Ranges for MELD and FIB-4 scores are 6–40 and 0.2–10, respectively, where lower scores represent milder disease and less fibrosis, respectively.
Disease progression was defined by changes in liver function tests and MELD and FIB-4 scores over time.
Patient selection
The existing database as of January 1, 2017, was searched for patients with a diagnosis of ALD in whom serum immunoglobulin levels had been measured at presentation or within 6 months thereafter. Patients with additional liver diseases were excluded, as were those with follow-up durations of less than 90 days. The remaining patients were classified as those with E-IgA or N-IgA levels (≥3.8 g/L or <3.8 g/L, respectively).
Serum immunoglobulin and other laboratory testing had been performed by the hospital’s clinical laboratories using commercial assays or standard techniques. This study was approved by the University of Manitoba’s human ethics review committee.
Categorical variables were evaluated using χ2 analysis (and F tests when warranted). Continuous variables were assessed using the Student t test or analysis of variance. Statistical significance was considered present when a p value fell below 5% in all analyses. The 95% CIs for significant differences were computed. Testing for correlations between serum IgA and the variables used to assess the severity of liver disease (bilirubin, INR, albumin, and MELD score) was performed. Cox proportional hazard models were developed for the risk factors associated with the development of cirrhosis over the length of the follow-up period as indicated by an increase of FIB-4 to more than 3.25 at the end of follow-up period among those whose FIB-4 was less than 3.25 at baseline. Stepwise backward selection of variables was used to fit the final model.
Results
Study populations
The initial database search identified 321 patients with ALD. Seventy-five were excluded because their serum IgA levels had not been tested within 6 months of presentation, 45 because they had been followed for less than 90 days, and 26 because they had coexisting liver disease (Figure 1). Thus, the total study population consisted of 175 patients with ALD.
Figure 1:
Derivation of study cohort
Note: The sum of individual diagnoses does not equal the total number of excluded patients as some patients had more than one coexisting liver disease
Baseline results
The gender distribution, mean age, blood cell counts, serum immunoglobulin levels, and liver enzyme tests at baseline are provided in Table 1. The majority of patients were male (65%). Mean age at presentation was 50 (SD 13) years. Mean hemoglobin, white blood cell, and platelet counts were all within the normal ranges. Mean serum AST, ALT, ALP, and GGT values were elevated, with AST exceeding ALT values (mean AST/ALT ratio 1.4, [SD 1.0]) and GGT being disproportionately elevated relative to ALP values (10.3-fold versus 1.2-fold, respectively). Mean serum IgA was elevated, but mean serum IgG and IgM levels were not.
Table 1:
Baseline demographic and laboratory values (N = 175)
| Characteristic | Normal range | All patients, mean (SD)* |
|---|---|---|
| Male, no. (%) | 113 (65) | |
| Age, y | 50 (13) | |
| Blood cell count | ||
| Hemoglobin, g/L | 140–180 | 130 (25) |
| WBC, × 109/L | 4.5–11.0 | 7.5 (3.7) |
| Platelets, × 109/L | 140–440 | 172 (80) |
| Liver enzymes, U/L | ||
| ALT | < 30 | 61 (72) |
| AST | 10–32 | 85 (92) |
| ALP | 30–120 | 146 (121) |
| GGT | 5–38 | 390 (507) |
| Immunoglobulins, g/L | ||
| IgA | 0.7–3.8 | 5.5 (3.2) |
| IgG | 6.9–16.2 | 15.0 (5.9) |
| IgM | 0.6–2.6 | 1.9 (1.4) |
* Unless otherwise indicated
WBC = white blood cells; ALT = alanine aminotransferase; AST = aspartate aminotransferase; ALP = alkaline phosphatase; GGT = gamma-glutamyl transferase; IgA = immunoglobulin A; IgG = immunoglobulin G; IgM = immunoglobulin M
The results of liver function tests and non- invasive indicators of disease severity for the entire study population are provided in Table 2. Mean serum bilirubin and INR, but not albumin levels, were abnormal relative to the normal laboratory ranges. The median MELD score was 9.6 (inter-quartile range [IQR] 7.5 to 14.4); the median FIB-4 score was 3.0 (IQR 1.8 to 5.7). Of the study participants 29 (17%) had FIB-4 scores compatible with no significant fibrosis (FIB-4 <1.45), and 80 (46%) had FIB-4 scores suggestive of advanced fibrosis or cirrhosis (FIB-4 >3.25). The remainder of the patients had intermediate FIB-4 values.
Table 2:
Baseline liver function tests and non-invasive indicators of disease severity (N = 175)
| Characteristic | Normal range | All patients, mean (SD)* |
|---|---|---|
| Liver function tests | ||
| Albumin, g/L | 33–45 | 35 (7) |
| Bilirubin, mmol/L | 5–21 | 39 (63) |
| INR | 0.9–1.1 | 1.3 (0.4) |
| Disease severity | ||
| MELD, Mdn (IQR) | 9.6 (7.5 to 14.4) | |
| FIB-4, Mdn (IQR) | 3.0 (1.8 to 5.7) | |
| FIB-4 < 1.45, no. (%) | 29 (17) | |
| FIB-4 > 3.25, no. (%) | 80 (46) |
* Unless otherwise indicated
INR = international normalized ratio; MELD = model for end-stage liver disease; FIB-4 = Fibrosis 4; IQR = inter-quartile range
Of the 175 study participants 115 (66%) had elevated serum IgA levels. As shown in Table 3, the majority of both E-IgA and N-IgA patients were male (67% and 60%, respectively). Their mean ages at diagnosis were also similar (mean 54 [SD 10] and mean 52 [SD 12] y respectively). E-IgA patients had lower mean hemoglobin levels than N-IgA patients (mean 126 [SD 26] and mean 137 [SD 21], respectively; p < .01) and platelet counts (mean 158 [SD 68] and mean 201 [SD 93], respectively; p < .001). Aside from serum ALT values, which were lower among E-IgA patients than N-IgA patients (mean 50 [SD 56] and mean 83 [SD 91] U/L, respectively; p < .005), liver enzyme levels were similar in the two study cohorts. As per the study protocol, serum IgA levels were higher among E-IgA than N-IgA patients (mean 7.1 [SD 2.9] and mean 2.5 [SD 0.9], respectively). The same was found for serum IgG values (mean 17.4 [SD 5.5] and mean 10.4 [SD 3.5] g/L respectively; p < .000001).
Table 3:
Baseline demographic and laboratory results in patients with elevated versus normal serum IgA levels
| Mean (SD)* | ||||
|---|---|---|---|---|
| Characteristic | Normal range | E-IgA (n = 115) | N-IgA (n = 60) | p |
| Male, n (%) | 77 (67) | 36 (60) | .36 | |
| Age at diagnosis, y | 54 (10) | 52 (12) | .34 | |
| Blood cell count | ||||
| Hemoglobin, g/L | 140–180 | 126 (26) | 137 (21) | .006 |
| WBC, × 109/L | 4.5–11.0 | 7.8 (4.1) | 7.1 (2.9) | .29 |
| Platelets, × 109/L | 140–440 | 158 (68) | 201 (93) | <.001 |
| Liver enzymes, U/L | ||||
| ALT | < 30 | 50 (56) | 83 (91) | .004 |
| AST | 10–32 | 84 (84) | 87 (106) | .85 |
| ALP | 30–120 | 157 (14) | 124 (56) | .081 |
| GGT | 5–38 | 394 (529) | 383 (464) | .89 |
| Immunoglobulins, g/L | ||||
| IgA | 0.7–3.8 | 7.1 (2.9) | 2.5 (0.9) | <.001 |
| IgG | 6.9–16.2 | 17.4 (5.5) | 10.4 (3.5) | <.001 |
| IgM | 0.6–2.6 | 2.1 (2.1) | 1.7 (1.2) | .24 |
| Liver function tests | ||||
| Albumin, g/L | 33–45 | 33 (7) | 38 (6) | <.001 |
| Bilirubin, mmol/L | 5–21 | 48 (74) | 22 (25) | .009 |
| INR | 0.9–1.1 | 1.3 (0.4) | 1.1 (0.2) | <.001 |
| Disease severity | ||||
| MELD, median (IQR) | 11.2 (8.3–15.5) | 7.7 (6.4–11.6) | <.001 | |
| FIB-4, median (IQR) | 3.6 (2.1–6.3) | 2.1 (1.2–3.7) | <.001 | |
| FIB-4 < 1.45, no. (%) | 10 (8.7) | 19 (32) | <.001 | |
| FIB-4 > 3.25, no. (%) | 63 (55) | 17 (28) | <.001 | |
* Unless otherwise indicated
IgA = immunoglobulin A; E-IgA = elevated IgA; N-IgA = normal IgA; WBC = white blood cells; ALT = alanine aminotransferase; AST = aspartate aminotransferase; ALP = alkaline phosphatase; GGT = gamma-glutamyl transferase; IgG = immunoglobulin G; IgM = immunoglobulin M; INR = international normalized ratio; MELD = model for end-stage liver disease; IQR = inter-quartile range; FIB-4 = Fibrosis 4
The results of liver function tests and MELD and FIB-4 scores for E-IgA and N-IgA patients are also provided in Table 3. E-IgA patients had significantly more hepatic dysfunction than N-IgA patients (lower albumin [mean 33 {SD 7} and mean 38 {SD 6} g/L, respectively], higher bilirubin [mean 48 {SD 74} and mean 22 {SD 25} μmol/L, respectively], and prolonged INR [mean 1.3 {SD 0.4} and mean 1.1 {SD 0.2}; ps < .01]). Median MELD scores were higher in the E-IgA cohort than in the N-IgA cohort (mean 11.2 [IQR 8.3 to 15.5] and mean 7.7 [IQR 6.4 to 11.6], respectively; p < .00001), as were FIB-4 scores (mean 3.6 [IQR 2.1 to 6.3] and mean 2.1 [IQR 1.2 to 3.7], respectively; p < .0001). E-IgA patients less frequently had FIB-4 scores compatible with minimal or no fibrosis than did N-IgA patients (FIB-4 < 1.45; 8.7% and 32%, respectively) and more frequently had FIB-4 scores in keeping with advanced fibrosis or cirrhosis (FIB-4 > 3.25; 55% and 28%, respectively; ps < .001).
Linear regression analysis was performed to determine whether correlations existed between serum IgA levels and the variables used to assess the severity of liver disease (albumin, bilirubin, INR, MELD, and FIB-4). The only significant correlation identified was that between serum IgA and MELD scores (r2 = .386, p < .001; Figure 2).
Figure 2:
Correlation coefficient for serum IgA levels and MELD scores at baseline (A) and last follow-up visit (B).
IgA = immunoglobulin A; MELD = model for end-stage liver disease
Because serum IgA levels can be elevated as a result of non-specific hypergammaglobulinemia associated with cirrhosis and portal-systemic shunting, the preceding analyses were repeated for ALD patients with normal serum IgG and IgM values (N = 86 patients). As shown in Table 4, among these individuals differences in liver function tests, MELD, FIB-4, and percentage of patients with FIB-4 scores of less than 1.45 and more than 3.25 remained significant (ps < .05).
Table 4:
Demographics, laboratory results, and non-invasive indicators of liver disease severity in adult ALD patients with elevated serum IgA but normal IgG and IgM levels
| Mean (SD)* | |||||
|---|---|---|---|---|---|
| Characteristic | Normal range | All patients (N = 86) | E-IgA (n = 43) | N-IgA (n = 43) | p |
| Age at diagnosis, y | 53.6 (10) | 53.5 (9) | 53.7 (11) | .94 | |
| Male, no. (%) | 56 (65) | 30 (70) | 26 (61) | .37 | |
| Blood cell count | |||||
| Hemoglobin, g/L | 140–180 | 139 (23) | 136 (25) | 142 (20) | .21 |
| WBC × 109/L | 4.5–11.0 | 7.1 (2.4) | 7.2 (2.0) | 7.0 (2.7) | .75 |
| Platelets × 109/L | 140–440 | 193 (81) | 176 (72) | 209 (87) | .06 |
| Liver enzymes, U/L | |||||
| ALT | < 30 | 67 (71) | 59 (63) | 76 (79) | .27 |
| AST | 10–32 | 76 (77) | 87 (84) | 65 (68) | .20 |
| AST/ ALT | 1.4 (0.8) | 1.7 (0.9) | 1.2 (0.7) | .08 | |
| ALP | 30–120 | 133 (94) | 147 (121) | 118 (54) | .15 |
| GGT | 5–38 | 381 (452) | 395 (458) | 367 (451) | .78 |
| Immunoglobulins, g/L | |||||
| IgA | 0.7–3.8 | 4.1 (2.9) | 5.6 (1.6) | 2.5 (0.8) | <.001 |
| IgG | 6.9–16.2 | 11.5 (2.6) | 12.9 (2.0) | 10.1 (2.6) | <.001 |
| IgM | 0.6–2.6 | 1.2 (0.4) | 1.23 (0.4) | 1.13 (0.5) | .29 |
| Liver function tests | |||||
| Albumin, g/L | 33–45 | 38 (6) | 37 (6) | 40 (5) | .02 |
| Total bilirubin, mmol/L | 5–21 | 26 (44) | 36 (58) | 17 (20) | .04 |
| INR | 0.9–1.1 | 1.1 (0.1) | 1.20 (0.2) | 1.07 (0.2) | .004 |
| Disease severity | |||||
| MELD, median (IQR) | 7.99 (6.43 to 11.97) | 9.55 (7.50 to 13.12) | 6.43 (6.43 to 8.77) | .016 | |
| FIB-4, median (IQR) | 2.17 (1.45 to 4.10) | 2.88 (1.80 to 6.00) | 1.69 (1.20 to 2.88) | .006 | |
| FIB-4 < 1.45, no. (%) | 22 (26) | 6 (14) | 16 (37) | .013 | |
| FIB-4 > 3.25, no. (%) | 25 (29) | 19 (44) | 6 (14) | .002 | |
* Unless otherwise indicated
ALD = alcohol-induced liver disease; IgA = immunoglobulin A; IgG = immunoglobulin G; IgM = immunoglobulin M; E-IgA = elevated IgA; N-IgA = normal IgA; WBC = white blood cells; ALT = alanine aminotransferase; AST = aspartate aminotransferase; ALP = alkaline phosphatase; GGT = gamma-glutamyltransferase; INR = international normalized ratio; MELD = model for end-stage liver disease; IQR = inter-quartile range; FIB-4 = Fibrosis 4
To confirm these findings, patients with FIB-4 scores of less than 3.25 were considered separately; once again, the results were similar (Table 5).
Table 5:
Demographics, laboratory results, and non-invasive indicators of liver disease severity in 95 adult ALD patients with FIB-4 scores < 3.25
| Mean (SD)* | |||||
|---|---|---|---|---|---|
| Characteristic | Normal range | All patients (N = 95) | E-IgA (n = 52) | N-IgA (n = 43) | p |
| Age at diagnosis, y | 52 (11) | 52 (11) | 52 (11) | .88 | |
| Male, no. (%) | 65 (68) | 38 (73) | 27 (63) | .28 | |
| Blood cell count | |||||
| Hemoglobin, g/L | 140–180 | 132 (26) | 124 (28) | 143 (18) | <.001 |
| WBC, × 109/L | 4.5–11.0 | 7.8 (4.3) | 8.2 (5.1) | 7.3 (2.8) | .33 |
| Platelets, × 109/L | 140–440 | 205 (75) | 193 (67) | 221 (83) | .07 |
| Liver enzymes, U/L | |||||
| ALT | < 30 | 58 (68) | 39 (43) | 81 (84) | .002 |
| AST | 10–32 | 50 (43) | 42 (27) | 59 (55) | .06 |
| AST/ALT | 1.25 (0.7) | 1.5 (0.8) | 1.0 (0.5) | <.001 | |
| ALP | 30–120 | 136 (101) | 155 (127) | 113 (48) | .05 |
| GGT | 5–38 | 309 (385) | 280 (331) | 345 (443) | .42 |
| Immunoglobulins, g/L | |||||
| IgA | 0.7–3.8 | 5.0 (3.1) | 7.0 (2.9) | 2.6 (0.9) | <.001 |
| IgG | 6.9–16.2 | 14.3 (5.7) | 17.3 (5.5) | 10.6 (3.1) | <.001 |
| IgM | 0.6–2.6 | 1.7 (1.2) | 1.8 (1.4) | 1.5 (1.0) | .19 |
| Liver function tests | |||||
| Albumin, g/L | 33–45 | 37 (7) | 33 (7) | 40 (5) | <.001 |
| Total bilirubin, mmol/L | 5–21 | 29 (64) | 43 (84) | 12 (11) | .02 |
| INR | 0.9–1.1 | 1.2 (0.3) | 1.3 (0.4) | 1.0 (0.1) | <.001 |
| Disease severity | |||||
| MELD, median (IQR) | 8.36 (6.43 to 11.63) | 9.61 (7.92 to 14.59) | 6.43 (6.43 to 8.30) | <.001 | |
| FIB-4, median (IQR) | 1.88 (1.15 to 2.00) | 2.21 (1.33 to 3.13) | 1.61 (1.04 to 2.76) | .03 | |
| FIB-4 < 1.45, no. (%) | 29 (31) | 10 (19) | 19 (44) | .009 | |
* Unless otherwise indicated
ALD = alcohol-induced liver disease; FIB-4 = Fibrosis 4; IgA = immunoglobulin A; E-IgA = elevated IgA; N-IgA = normal IgA; WBC = white blood cells; ALT = alanine aminotransferase; AST = aspartate aminotransferase; ALP = alkaline phosphatase; GGT = gamma-glutamyltransferase; IgG = immunoglobulin G; IgM = immunoglobulin M; INR = international normalized ratio; MELD = model for end-stage liver disease; IQR = inter-quartile range
Follow-Up
The mean durations of follow-up for E-IgA and N-IgA patients were in excess of 50 months with no difference between groups (mean 54 [SD 46] mo and mean 51 [SD 47] mo, respectively). At the end of follow-up, E-IgA patients had lower platelet counts, similar liver enzyme levels, and higher serum IgA and IgG levels when compared with N-IgA patients (Table 6).
Table 6:
Demographics, laboratory results, and non-invasive indicators of liver disease severity in 175 adult ALD patients with and without elevated serum IgA levels at the end of follow-up
| Mean (SD)* | |||||
|---|---|---|---|---|---|
| Characteristic | Normal range | All patients (N = 175) | E-IgA (n = 115) | N-IgA (n = 60) | p |
| Follow-up, mo | 53 (46) | 54 (46) | 51 (47) | .98 | |
| Blood cell count | |||||
| Hemoglobin, g/L | 140–180 | 131 (24) | 129 (24) | 135 (22) | .14 |
| WBC, × 109/L | 4.5–11.0 | 6.6 (2.4) | 6.5 (2.5) | 6.8 (2.2) | .34 |
| Platelets, × 109/L | 140–440 | 170 (84) | 160 (89) | 189 (71) | .03 |
| Liver enzymes, U/L | |||||
| ALT | < 30 | 45 (44) | 48 (50) | 39 (27) | .24 |
| AST | 10–32 | 34 (34) | 32 (32) | 39 (38) | .20 |
| AST/ALT | 1.5 (0.8) | 1.6 (0.7) | 1.3 (0.8) | .08 | |
| ALP | 30–120 | 120 (76) | 124 (81) | 114 (67) | .43 |
| GGT | 5–38 | 196 (305) | 193 (305) | 202 (306) | .86 |
| Immunoglobulins, g/L | |||||
| IgA | 0.7–3.8 | 4.1 (2.0) | 6.8 (2.9) | 2.8 (1.3) | <.001 |
| IgG | 6.9–16.2 | 11.5 (2.6) | 16.7 (5.5) | 10.9 (3.4) | <.001 |
| IgM | 0.6–2.6 | 1.2 (0.4) | 2.1 (2.1) | 1.7 (1.2) | .24 |
| Liver function tests | |||||
| Albumin, g/L | 33–45 | 36 (6) | 35 (6) | 38 (6) | <.001 |
| Total bilirubin, mmol/L | 5–21 | 22 (39) | 25 (41) | 18 (33) | .28 |
| INR | 0.9–1.1 | 1.2 (0.3) | 1.2 (0.3) | 1.1 (0.3) | .004 |
| Disease severity | |||||
| MELD, median (IQR) | 6–40 | 8.5 (6.5 to 12.9) | 9.4 (7.5 to 13.5) | 7.2 (6.4 to 9.4) | <.001 |
| FIB-4, median (IQR) | 2.3 (1.4 to 4.1) | 2.6 (1.5 to 5.5) | 1.8 (1.2 to 3.2) | .001 | |
| FIB-4 < 1.45, no. (%) | 47 (27) | 24 (21) | 23 (38) | .013 | |
| FIB-4 > 3.25, no. (%) | 60 (34) | 46 (40) | 14 (23) | .03 | |
* Unless otherwise indicated
ALD = alcohol-induced liver disease; IgA = immunoglobulin A; E-IgA = elevated IgA; N-IgA = normal IgA; WBC = white blood cells; ALT = alanine aminotransferase; AST = aspartate aminotransferase; ALP = alkaline phosphatase; GGT = gamma-glutamyltransferase; IgG = immunoglobulin G; IgM = immunoglobulin M; INR = international normalized ratio; MELD = model for end-stage liver disease; IQR = inter-quartile range; FIB-4 = Fibrosis 4
With respect to liver function, aside from mean serum bilirubin levels, which were now similar in the two cohorts, albumin levels continued to be lower (mean 35 [SD 6] and mean 38 [SD 6], respectively) and INR values continued to be higher (mean 1.2 [SD 0.3] and mean 1.1 [SD 0.3], respectively) in E-IgA patients than in N-IgA patients (p < .005). Similarly, median MELD (median 9.4 [IQR 7.5 to 13.5] and median 7.2 [IQR 6.4 to 9.4], respectively) and FIB-4 (median 2.6 [IQR 1.5 to 5.5] and median 1.8 [IQR 1.2 to 3.2], respectively) scores continued to be higher in the E-IgA cohort than in the N-IgA cohort (p < .001). The percentage of individuals with FIB-4 values less than 1.45 and more than 3.25 were 21% versus 38% and 40% versus 23% in the E-IgA and N-IgA cohorts, respectively (p < .05). These findings were similar when E-IgA patients with normal serum IgG and IgM levels and FIB-4 scores of less than 3.25 were considered separately at the end of follow-up (data not shown).
Progression
Table 7 provides the results of changes from baseline in liver enzyme and function tests and MELD and FIB-4 scores at the end of follow-up. With the exception of serum ALT values, which declined to a lesser extent in E-IgA patients than in N-IgA patients (mean 18 [SD 50] and mean 44 [SD 93] U/L, respectively; p < 0.05), liver enzymes and the AST/ALT ratio declined to similar extents in both cohorts. Changes in liver function tests (albumin, bilirubin, and INR) were also similar in the two cohorts. Although the decline in median MELD and FIB-4 scores and percentage of patients with FIB-4 values greater than 3.25 were similar in the two cohorts, significantly more E-IgA patients with initial FIB-4 scores greater than 3.25 had scores decline to less than 3.25 than did N-IgA patients (15% versus 5%, p = 0.05). These findings at the end of follow-up were similar when E-IgA patients with normal serum IgG and IgM levels and those with FIB-4 scores less than 3.25 were considered separately (data not shown).
Table 7:
Changes in liver enzymes, liver function, MELD, and FIB-4 scores from baseline to end of follow-up in 175 adult ALD patients with and without elevated serum IgA levels
| Mean (SD)* | ||||
|---|---|---|---|---|
| Characteristic | All patients (N = 175) | E-IgA (n = 115) | N-IgA (n = 60) | p |
| Liver enzymes, U/L | ||||
| ΔAST | −2.8 (146) | −35 (177) | −16 (40) | .44 |
| ΔALT | −27 (69) | −18 (50) | −44 (93) | .02 |
| ΔAST/ALT | −0.3 (1.0) | −0.4 (0.9) | −0.1 (1.2) | .11 |
| ΔALP | −26 (111) | −34 (129) | −10 (63) | .17 |
| ΔGGT | −88 (411) | −91 (486) | −82 (202) | .89 |
| Liver function | ||||
| ΔAlbumin, g/L | 1.4 (8) | 2 (8) | 0 (7) | .12 |
| ΔTotal bilirubin, µmol/L | −17 (69) | −23 (82) | −4 (27) | .08 |
| ΔINR | −0.1 (0.4) | −0.1 (0.4) | −0.0 (0.2) | .17 |
| XXX | ||||
| ΔIgA, g/L | −0.1 (0.4) | −0.32 (2.2) | 0.28 (1.3) | .03 |
| Disease severity | ||||
| ΔMELD | −1.1 (5.5) | −1.6 (6.2) | −0.1 (3.9) | .08 |
| ΔFIB-4 score | −1.7 (10.8) | −2.1 (12.8) | −1.0 (5.2) | .54 |
| Developed FIB-4 <1.45, no. (%) | 18 (10) | 14 (12) | 4 (7) | .23 |
| Developed FIB-4 >3.25, no. (%) | 20 (11) | 11 (10) | 3 (5) | .054 |
* Unless otherwise indicated
MELD = model for end-stage liver disease; FIB-4 = Fibrosis 4; ALD = alcohol-induced liver disease; IgA = immunoglobulin A; E-IgA = elevated IgA; N-IgA = normal IgA; AST = aspartate aminotransferase; ALT = alanine aminotransferase; ALP = alkaline phosphatase; GGT = gamma-glutamyltransferase; INR = international normalized ratio
Cox proportional hazard models for the risk factors associated with the development of cirrhosis (FIB-4 > 3.25) at the end of follow-up were developed (Table 8). Only age at diagnosis and an initial FIB-4 level of less than 1.45 predicted the development of cirrhosis (HR 1.10 [95% CI 1.04 to 1.16] and HR 0.51 [95% CI 0.33 to 0.79], respectively).
Table 8:
Cox proportional hazard model for predicting cirrhosis (FIB-4 < 3.25) in non-cirrhotic ALD patients
| Variable | Hazard ratio (95% CI) | p |
|---|---|---|
| Age at diagnosis, y | 1.0950 (1.0356 to 1.1597) | .0015 |
| Baseline FIB-4 <1.45 | 0.5086 (0.3265 to 0.7922) | .0028 |
FIB-4 = Fibrosis-4; ALD = alcohol-induced liver disease
After a similar period between testing, changes in serum IgA levels were significant in the E-IgA but not the N-IgA cohort when compared with baseline values (mean 0.32 [SD 2.2] and mean 2.28 [SD 1.3], respectively; p = .03). At the end of follow-up, 95% of E-IgA patients continued to have elevated levels and 93% of N-IgA patients continued to have normal levels.
Discussion
The results of this study indicate that the majority of ALD patients had elevated serum IgA levels. When present, this finding was associated with non-invasive indicators of advanced liver disease. However, progression of the disease appeared to be independent of serum IgA values.
Previously published reports have described increased serum IgA levels in 50%–95% of adult ALD patients (12,13,16). Thus, our finding of 67% is well within the published range. It is important to note that increased serum IgA levels have also been described in individuals who drink alcohol to excess but have no evidence of liver disease (17,18). This finding further supports the contention that IgA does not contribute to the pathogenesis of ALD but does not preclude a protective effect.
The explanation for why serum IgA levels are elevated in the majority of ALD patients remains to be determined. Proposed theories include excess IgA synthesis and secretion by plasma cells within the Peyer’s patches of the intestine in response to gut-derived immunogens or decreased hepatic clearance due to portal-systemic shunting or reticuloendothelial system (RES) dysfunction (19–22). Although our finding that serum IgA levels are associated with biochemical evidence of liver dysfunction would support portal-systemic shunting or decreased hepatic/RES clearance, the fact that elevated levels were also observed in patients with minimal fibrosis (low FIB-4 scores) argues against that explanation.
Several means of documenting the severity of liver disease in ALD are available. Because of the limited number of biopsies performed among this patient population, we relied on standard tests of liver function and MELD and FIB-4 scores as surrogate markers. Each variable has been tested and validated to be of value in predicting the extent of ALD (23,24). Nonetheless, the results of liver histology would have been of interest, particularly in determining the extent of active liver disease because serum aminotransferase, ALP, and GGT levels are of limited value in that regard.
The ability to predict the course of ALD is becoming of increasing clinical importance as therapeutic interventions are identified and new agents proceed through clinical trials (3,25). Unfortunately, our data do not support the use of serum IgA levels for this purpose. Thus, other biomarkers will need to be identified and tested, preferably in prospective controlled trials.
This study has several limitations that warrant emphasis. First, the study was retrospective and limited to a single, tertiary-care referral centre. Second, an additional selection bias may have been introduced in that not all ALD patients had serum immunoglobulin levels documented within 6 months of their initial visit. Third, liver histology was not available to document liver tissue IgA deposition or confirm the presence or absence of advanced fibrosis or cirrhosis. Fourth, the amount of alcohol consumed by patients throughout the course of follow-up was not documented. Finally, the impact of therapeutic interventions beyond abstinence such as corticosteroids and antibiotics that might alter the severity or course of ALD or serum IgA levels was not ascertained.
In conclusion, the results of this study indicate that the majority of adult ALD patients have elevated serum IgA levels, but baseline levels are not predictive of the subsequent course. Hence, IgA is unlikely to play an important role in the pathogenesis of ALD.
Acknowledgements:
The authors thank R. Vizniak for her prompt and accurate typing of the manuscript.
Ethics Approval:
The study protocol was approved by an ethics committee and the ethics certificate information is available from the authors upon request.
Informed Consent:
Informed consent was obtained from the patients.
Registry and Registration No. of The Study/ Trial:
N/A
Funding:
None to declare.
Disclosures:
None to declare.
Peer Review:
This article has been peer reviewed.
References
- 1.National Institutes of Health. Alcohol use and alcohol use disorders in the United States: main findings from the 2001-2002 National Epidemiologic Survey on Alcohol and Related Conditions. US Alcohol Epidemiologic Data Reference Manual. 2006;8(1). [Google Scholar]
- 2.Yoon YH, Chen CM. Liver cirrhosis mortality in the United States: national, state, and regional trends, 2000-2013. National Institute on Alcohol Abuse and Alcoholism Surveillance Report No. 105. Bethesda: National Institute on Alcohol Abuse and Alcoholism; 2016. [Google Scholar]
- 3.Louvet A, Mathurin P. Alcoholic liver disease: mechanisms of injury and targeted treatment. Nat Rev Gastroenterol Hepatol. 2015;12:4:231–42. 10.1038/nrgastro.2015.35. Medline: [DOI] [PubMed] [Google Scholar]
- 4.Bode C, Kugler V, Bode JC. Endotoxemia in patients with alcoholic and non-alcoholic cirrhosis and in subjects with no evidence of chronic liver disease following acute alcohol excess. J Hepatol. 1987;4(1):8–14. 10.1016/S0168-8278(87)80003-X. [DOI] [PubMed] [Google Scholar]
- 5.Mathurin P, Deng QG, Keshavarzian A, Choudhary S, Holmes EW, Tsukamoto H. Exacerbation of alcoholic liver injury by enteral endotoxin in rats. Hepatology. 2000; 32(5):1008–17. 10.1053/jhep.2000.19621. Medline: [DOI] [PubMed] [Google Scholar]
- 6.Parlesak A, Schafer C, Schutz T, Bode JC, Bode C. Increased intestinal permeability to macromolecules and endotoxemia in patients with chronic alcohol abuse in different stages of alcohol-induced liver disease. J Hepatol. 2000;32(5):742–7. 10.1016/S0168-8278(00)80242-1. [DOI] [PubMed] [Google Scholar]
- 7.Tamai H, Horie Y, Kato S, Yokoyama H, Ishii H. Long-term ethanol feeding enhances susceptibility of the liver to orally administered lipopolysaccharides in rats. Alcohol Clin Exp Res. 2002;26(8 Suppl):75S–80S. 10.1097/00000374-200208001-00016. [DOI] [PubMed] [Google Scholar]
- 8.Hanck C, Rossol S, Bocker U, Tokus M, Singer MV. Presence of plasma endotoxin is correlated with tumor necrosis factor receptor levels and disease activity in alcoholic cirrhosis. Alcohol Alcohol. 1998;33(6):606–8. 10.1093/alcalc/33.6.606. Medline: [DOI] [PubMed] [Google Scholar]
- 9.Fujimoto M, Uemura M, Nakatani Y, et al. Plasma endotoxin and serum cytokine levels in patients with alcoholic hepatitis: relation to severity of liver disturbance. Alcohol Clin Exp Res. 2000;24(4 Suppl):48S–54S. 10.1111/j.1530-0277.2000.tb00012.x. [DOI] [PubMed] [Google Scholar]
- 10.Corthesy B. Role of secretory IgA in infection and maintenance of homeostasis. Autoimmun Rev. 2013;12(6)661–5. 10.1016/j.autrev.2012.10.012. Medline: [DOI] [PubMed] [Google Scholar]
- 11.Brown WR, Kloppel TM. The liver and IgA: immunological, cell biological and clinical implications. Hepatology. 1989;9(5):763–84. 10.1002/hep.1840090518. Medline: [DOI] [PubMed] [Google Scholar]
- 12.van de Wiel A, Delacroix D, van Hattum J, Schuurman HJ, Kater L. Characteristics of serum IgA and liver IgA deposits in alcoholic liver disease. Hepatology. 1987;7(1):95–9. 10.1002/hep.1840070120. Medline: [DOI] [PubMed] [Google Scholar]
- 13.van de Wiel A, van Hattum J, Schuurman HJ, Kater L. Immunoglobulin A in the diagnosis of alcoholic liver disease. Gastroenterology. 1988;94(2):457–62. 10.1016/0016-5085(88)90437-4. [DOI] [PubMed] [Google Scholar]
- 14.Bogers WM, Stad RK, van Es LA, Daha MR. Immunoglobulin A: interaction with complement, phagocytic cells and endothelial cells. Complement Inflamm. 1991;8(5-6): 347–58. 10.1159/000463206. Medline: [DOI] [PubMed] [Google Scholar]
- 15.Parlesak A, Schafer C, Bode C. IgA against gut-derived endotoxins: does it contribute to suppression of hepatic inflammation in alcohol-induced liver disease? Dig Dis Sci. 2002;47(4):760–66. 10.1023/A:1014783815433. Medline: [DOI] [PubMed] [Google Scholar]
- 16.Lee FI, Lond MB. Immunoglobulins in viral hepatitis and active alcoholic liver disease. Lancet. 1965;1043–1046. 10.1016/S0140-6736(65)90571-4. [DOI] [PubMed] [Google Scholar]
- 17.Iturriaga H, Pereda T, Estevez A, Ugarte G. Serum immunoglobulin A changes in alcoholic patients. Ann Clin Res. 1977;9(1): 39–43. [PubMed] [Google Scholar]
- 18.Gonzalez-Quintela A, Alende MR, Gamallo R et al. Serum immunoglobulins (IgG, IgA, IgM) in chronic hepatitis C. A comparison with non-cirrhotic alcoholic liver disease. Hepatogastroenteroly. 2003;50(54):2121–6. [PubMed] [Google Scholar]
- 19.van de Wiel A, Seifert WF, Van der Linden JA, Gmelig-Meyling FH, Kater L, Schuurman HJ. Spontaneous IgA synthesis by blood mononuclear cells in alcoholic liver disease. Scand J Immunol. 1987;25(2):181–7. 10.1111/j.1365-3083.1987.tb01062.x. Medline: [DOI] [PubMed] [Google Scholar]
- 20.Allison ME, Hodgson HJ. Regulation of peripheral blood B-cell IgA production in alcoholic cirrhosis. J Clin Lab Immunol. 1989;30(3):127–30. [PubMed] [Google Scholar]
- 21.Calmus Y, Poupon R, Barbare JC, Johanet C, Nedelec P, Darnis F. Hepatic function and significance of the increase in serum concentrations of IgA in alcoholic cirrhosis. Gastroenterol Clin Biol. 1985;9(8-9):614–6. [PubMed] [Google Scholar]
- 22.Gomez F, Ruiz P, Schreiber AD. Impaired function of macrophage Fc gamma receptors and bacterial infection in alcoholic cirrhosis. N Engl J Med. 1994;331(17):1122–8. 10.1056/NEJM199410273311704. Medline: [DOI] [PubMed] [Google Scholar]
- 23.Naveau S, Gaudé G, Asnacios A, et al. Diagnostic and prognostic values of noninvasive biomarkers of fibrosis in patients with alcoholic liver disease. Hepatology. 2009;49(1):97–105. 10.1002/hep.22576. Medline: [DOI] [PubMed] [Google Scholar]
- 24.Fernandez M, Trepo E, Degre D, et al. Transient elastography using Fibroscan is the most reliable noninvasive method for the diagnosis of advance fibrosis and cirrhosis in alcoholic liver disease. Eur J Gastroenterol Hepatol. 2015;27(9):1074–9. 10.1097/MEG.0000000000000392. Medline: [DOI] [PubMed] [Google Scholar]
- 25.Thompson J, Jones N, Al-Khafaji A, et al. Extracorporeal cellular therapy (ELAD) in severe alcoholic hepatitis: a multinational, prospective, controlled, randomized trial. Liver Transpl. 2018;24(3):380–93. [DOI] [PMC free article] [PubMed] [Google Scholar]