Article Text
Abstract
Introduction We have previously proposed a linear relationship between the degree and duration of suppression of 24-h intragastric acidity and the healing rate of erosive esophagitis (EE) with the best predictor as per cent time pH ≥4. Poor response to PPI is an increasingly important concern. Simply reporting time pH <4 may not reflect the extent of acid exposure because of the log scale of pH (1 unit pH change is a 10-fold change in [H+ ion] and a 100-fold increase if pH falls from 4 to 2). Indeed, in healthy volunteers ∼40% of the 24-h is pH ≤4 after 5 days esomeprazole 40 mg om and pH is <2 for 42% of the time between midnight and 0700.1 We aimed to investigate the relationship of time pH <4 in the 24-h and the EE healing rate at 4 and 8 weeks with antisecretory drugs, since time pH <4 is the only widely available pH parameter.
Methods A meta-analysis was performed to provide 24-h intragastric pH values at steady state (day 5–8) in healthy volunteers and the EE healing rate at 4 or 8 weeks, by searching the English-language literature up to June 2009 in MEDLINE and EMBASE. For pH data, a weighted mean is only possible by using the sample size of each arm ((mean×n from each study)/Σ treated N). A regimen was excluded if <20 subjects provided pH data. For EE data, the pooled healing rate was calculated by weighting the sample size (random effect model) using StatsDirect®2.6.3. Here also a regimen was excluded if <30 subjects per arm provided EE data. Standard linear regression was performed by SPSS® 17.0, with the pooled EE healing rate at 4 weeks or 8 weeks as the dependent variable and mean per cent time pH <4 in 24-h of the corresponding treatment regimen as the independent variable.
Result 95 studies (411 arms) provided intragastric pH data and 109 studies (243 arms) provided EE healing data for 3 H2-RAs, 5 PPIs and placebo. The commonly reported pH parameter in the pH-EE model was mean per cent time pH <4, which was used in the model. At 4 weeks, the model for the EE (all grades) healing rate = 110.52−1.01* mean per cent time pH <4 in 24 h (R=0.862, p<0.001). At 8 weeks. The model for EE (all grades) healing rate=124.88−1.02* mean per cent time pH <4 in 24 h (R=0.887, p<0.001).
Conclusion Any increase in the time intragastric pH is pH <4 in the 24 h at steady state in healthy volunteers is associated with a decrease in the EE healing rate at 4 and 8 weeks. The per cent time that intragastric pH <4 is a significant predictor for EE non-healing. Although, not usually reported, a better predictor is likely to be the time pH is <3 or better pH <2, especially during the night-time period.