A randomized, double-blind, and controlled design was adopted in this trial. The physicians who were responsible for endoscopy, 24-hour esophageal pH monitoring, and administering the questionnaire survey were relatively independent. After tests, another physician performed the statistical analysis in order to guarantee the objectivity and validity of tests.
Results showed that the 90 patients consisted of 59 males and 31 females, with the male-to-female ratio approaching 2
1. This ratio is similar to previous reports and may be due to histories of smoking, and drinking. Reportedly [17
], drinking, smoking, obesity and overeating are major risk factors for GERD. No significant differences were noted in age, LESP, UESP, and SI among the three groups (P
> 0.05), suggesting that patients in various groups were comparable after randomized and double-blinded grouping.
Group A (40
mg/day for two weeks), Group B (20
mg/day for two weeks), and Group C (placebo for the 1st week and rabeprazole 20
mg/d for the 2nd week) were designed to investigate the effects of PPI in different doses and treatment duration on test results. Results indicate that SF-36 scores had no significant differences between pretreatment and after one or two weeks of treatment, which may be attributed to non-GERD patients mingling between all the groups. Therefore, GERD and non-GERD patients should be analyzed separately. When patients were grouped according to GERD and non-GERD diagnosis, differences in SF-36 scores were only noted between GERD and non-GERD groups after two weeks of treatment for Group A. The differences were not significant between GERD and non-GERD groups before and after one and two weeks of treatment in other groups. It is believed that SF-36 scores, the common disease scale, are affected not only by GERD itself but also the occupation, material status, family and social relationships, education, household income, and social class of patients. The Az was 0.27 in the SF-36 score ROC prior to the treatment, suggesting that SF-36 alone is a poor tool for diagnosing GERD in a primary care setting and is not sufficient to establish a GERD diagnosis. Further analysis showed that the differences were significant in improved SF-36 scores between GERD and non-GERD patients in Groups A and B (P
< 0.05) but not significant between GERD and non-GERD patients in Group C (P
= 0.085) after 1-week therapy; significant differences were noted between GERD and non-GERD patients among the three groups after two-week therapy (P
< 0.05). However, statistical differences in improvement rate were not noted between GERD and non-GERD patients in Groups B and C after two-week treatment (20
mg/d), as the improvement rate is correlated with improved scores and basic scores, and the basic SF-36 scores are related not only to GERD itself but also to many other factors mentioned above. In Group A (40
mg/d), statistical differences in improvement rate are noted between GERD and non-GERD patients, which may be strongly related to the increase of improved scores. Improved scores and basic scores can both influence the improvement rate. Therefore, improvement rate is not regarded as a criterion of the PPI test in improving scores. An improved score, the difference in life quality before and after the treatment, represents the degree of improvement in life quality and can be affected only by a few factors. Therefore, the improved score can be taken as a criterion for the PPI test. Our results show that according to the Youden value principle, an improved score to 65 is most efficient for a GERD diagnosis. At this level, the diagnostic sensitivity was 94.4% and the specificity was 78.9%. If the improved score is boosted, the specificity increases but the sensitivity decreases; therefore, an SF-36 score increase of 65 is taken as the criterion for a positive PPI test result.
In this trial, the diagnostic value of the rabeprazole test did not differ significantly according to duration; the diagnostic value can be considered approximately equivalent after one week of treatment and after two weeks of treatment. The diagnostic coincident rate is 73%, the sensitivity is 85.7%, and the specificity is 44.5% after a one-week administration of rabeprazole 10
mg twice per day. Significant differences in diagnostic sensitivity and specificity at the two-week mark are not noted compared to that at the one-week mark. Diagnostic efficacy is consistent and the expense increases significantly. Therefore, diagnostic administration for two weeks is unnecessary for judging results. There are no significant differences in the sensitivity and specificity of the 40
mg/day and 20
mg/day groups. Therefore, rabeprazole given as 10
mg b.i.d. for one week is optimum for the PPI test and has the added benefit of being less costly. Results of diagnostic tests with rabeprazole, as reported by Schenk et al. [18
], show that the sensitivity, specificity, and negative predictive value were, respectively, 68%, 63%, and 68% in the trial group and 20%, 95%, and 83% in the control group. Johnsson et al. [19
] conducted a trial with omeprazole (20
mg b.i.d.) and results showed that the one-week sensitivity was 75% and the specificity was 55%. Cho et al. [20
] reported that if lansoprazole 30
mg was given as b.i.d. for two weeks, the diagnostic sensitivity and specificity were 77% and 56%, respectively, illustrating that as the diagnostic sensitivity and specificity increased, the total coincident rate was similar, but the expense rose significantly. Therefore, the two-week administration did little to improve diagnosis. If patients with reflux symptoms are given rabeprazole 10
mg b.i.d. for one week and their SF-36 score increases 65 units following treatment, they can be diagnosed as GERD. Further logistic regression analysis suggests that the diagnostic sensitivity was 94.4%, specificity was 78.9%, the coincident rate was 91.1%, the false negative rate was 5.6%, and the false positive rate was 21.1%.
The SF-36 consists of 3 major parts: functional status, health satisfaction, and total evaluation. It includes eight fields: physical function, physical responsibility, body pain, activation, social function, and emotional responsibility. The eight fields are classified further into physical component scales and mental component scales. The SF-36, a common scale, not only measures its own items but also investigates several specific problems as affecting factors when determining the quality of life with GERD. It comprises more contents than the relatively limited RDQ. Some GERD patients present primarily with extraesophageal symptoms such as coughing and throat discomfort, which strongly influence the quality of life. After treatment, patients improved and their life quality increased. Therefore, rabeprazole in combination with the SF-36 can make a diagnosis through a comparison of pretreatment and posttreatment scores. A common scale may be more helpful than a GERD-specific scale to clarify the reason for decline in the quality of life.
The PPI test is a current diagnostic method for GERD. This study aims to determine the effectiveness of coapplying the PPI test and the SF-36 for GERD diagnosis. Rabeprazole is metabolized in nonenzymic fashion, with a longer half-life, more stable pharmacokinetics, and greater efficacy than the first generation of PPI. In clinical practice, administration of rabeprazole can improve reflux symptoms and the quality of life rapidly [21
]. In this study, SF-36 scores increase significantly after administration of rabeprazole and differences are more significant than the pretreatment. These differences are induced by rabeprazole for an individual with the same specific problems and thus interference from specific problems can be excluded. Administration of the basic SF-36 seems to have no value for diagnosing GERD at the patient's initial visit. However, increase in the SF-36 score after rabeprazole treatment can be used for the cut-off value in the rabeprazole test and thus provides the preliminary quantitative criteria for the PPI test. Large-sample, multicenter trials are required to confirm this result in clinical practice.
In conclusion, our study shows that the SF-36 in combination with the rabeprazole test can screen GERD patients and increase the sensitivity and specificity of GERD diagnosis through reference to the change in SF-36 score before and after the treatment (65 in the trial). This not only reduces the expense of clinical diagnosis but also reduces the pain that might be inflicted for gastroscopy and pH monitoring. With this method, diagnosis and treatment can be performed concurrently to shorten diagnosis duration.
Certainly, more exact diagnostic criteria require more large-sample, multicenter, randomized control, and double-blinded studies.