We found 29 candidates who met the screening criteria and were taken biopsy specimens during the study period. One patient was excluded from the study because the tissue from the corpus was too small to evaluate atrophy of fundic glands. Of the remaining 28 patients, 22 patients (78.6%, 95% confidence interval 59.0–91.7) were histopathologically confirmed AIG and studied in the following sections.
Of the 22 AIG patients, 7 were men and 15 were women. Age ranged 43–81 years (mean ± SD = 65.2 ± 10.8 years). Two patients used vitamin B12-containing agents. There were no patients using proton pump inhibitor (PPI). Currently H. pylori-infected patients were not found in any of the 22 AIG patients. On the other hand, 12 (54.5%) and 10 patients (45.5%) were diagnosed as past and not infected, respectively. Four patients (18.2%) and 18 patients (81.8%) were histopathologically classified into florid and end stages of AIG, respectively (Table 1, Fig. 1). No patients were histopathologically classified into early stage of AIG.
In 22 AIG patients, anti-PC antibody ranged from 10x to more than 160x (10x in 2 patients, 20x in 6 patients, 40x in 5 patients, 80x in 5 patients, 160x or greater in 4 patients). One patient was also positive for anti-IF antibody.
Serum gastrin was measured in all 22 AIG patients. In these AIG patients, gastrin ranged from 690 to 5300 (mean ± SD = 2218.2 ± 1509.6) pg/mL and all these values exceeded the normal range (172 pg/mL). Of these 22 patients, 4 and 18 patients were included in the florid and end stages, respectively. Serum gastrin was not statistically different between histopathological stages of AIG (florid stage: mean ± SD = 1657.5 ± 362.5 pg/mL; end stage: 2342.8 ± 1634.3 pg/mL; P = 0.14, t-test), but those who showed higher serum gastrin than 3000 pg/mL were all in end stage (Fig. 2).
We analyzed cut-off values for PGs with receiver operating characteristic (ROC) curves. The areas under the curve (AUC) of PG I, PG II and PG I/II ratio were 0.81, 0.29 and 0.98, respectively (Fig. 3), indicating that PG I and PG I/II ratio were suitable for making cut-off values. The optimal cut-off values for PG I and PG I/II ratio were suggested as 14.5 ng/mL (sensitivity = 0.83, specificity = 0.85) and 2.1 (sensitivity = 1.00, specificity = 0.95), respectively. We also evaluated Miki’s criteria for PG test in diagnosing AIG (Table 2). Sensitivity and specificity of PG 3+ were 85 and 100%, respectively. On the other hand, sensitivity and specificity of PG 2+ or 1+ were 90 and 66.7%, respectively. Among histopathologically confirmed AIG patients, PG I ranged from 2.5 to 71.9 (mean ± SD = 13.3 ± 19.5) ng/mL, PG II from 3.7 to 39.2 (mean ± SD = 10.9 ± 8.3) ng/mL and PG I/II ratio from 0.3 to 3.5 (mean ± SD = 1.0 ± 0.7). PG I, PG II and PG I/II ratio were not statistically different between histopathological stages (Table 3). PG test classification revealed 3+ in 17 (85%), 2+ in 1 (5%), 1+ in 0 (0%) and negative in 2 patients (10%) (Fig. 4). Of these 20 patients, 4 and 16 patients were included in florid and end stages, respectively, but the results of PG test were not related to the stage of AIG.
Serum vitamin B12 was measured in all 22 patients but 2 of them used vitamin B12-containing agents and the values were higher than 400 pg/mL. In the rest of 20 patients, vitamin B12 ranged from 50 to 320 (mean ± SD = 137.5 ± 79.2) pg/mL (Fig. 5). According to the cut-off value (180 pg/mL), 13 patients (65%) showed low serum vitamin B12. Of the 20 patients, 3 and 17 patients were included in florid and end stages, respectively (Fig. 5). Serum vitamin B12 was not statistically different between histopathological stages of AIG (florid stage: mean ± SD = 147.3 ± 37.8 pg/mL; end stage: 135.7 ± 84.3 pg/mL; P = 0.83, t-test). All 22 patients showed high serum gastrin and 20 of them did not use vitamin B12-containing agents. Among them, 7 and 13 patients showed normal and low serum vitamin B12, respectively (Fig. 6). On the other hand, in the 20 patients without vitamin B12 users, 7 and 13 showed normal and low serum vitamin B12, respectively. All the patients with low serum vitamin B12 showed high serum gastrin (Fig. 6).
Endoscopic corpus atrophy was evaluated in all the 22 AIG patients. Among them, 20 (90.9%), 1 (4.5%) and 1 (4.5%) were classified in O-p, O-3 and O-1, respectively. Therefore, all the 22 AIG patients had endoscopic atrophy O-1 or greater. Of 4 patients in florid stage, 3 were classified in O-p and one in O-3. Of 18 patients in end stage, 17 were classified in O-p and one in O-1. There was no significant difference in O-p ratio between florid and end stage (75% vs 94.4%, respectively, P = 0.22, chi-square test, Table 4).
We counted the number of positive items among the 4 tests: elevated serum gastrin, endoscopic O-p atrophy, 3+ PG test and low serum vitamin B12. Four items were positive only in 11 AIG patients (50%). On the other hand, the number of positive items was 1 in 1 patient, 2 in 3 patients and 3 in 7 patients, respectively (Table 1). Of the 7 patients whose number of positive items was 3, 6 patients had normal serum vitamin B12 and 1 had 2+ PG test. Common clinical findings in the confirmed AIG patients were 10x or greater anti-PC antibody, elevated serum gastrin greater than 172 pg/mL and endoscopic atrophy O-1 or greater. Other items were not common in all the 22 AIG patients.