Primary Motility  Disorders of the  Esophagus
 The Esophageal
 Esophagogastric  Junction

  Browse by Author
  Browse by Movies
Volume: Barrett's Esophagus
Chapter: Pathophysiology

Can quantitative data on alkaline refluxate be expected from any measurement system?

P. Burdiles (Santiago)

Quantitative data in medicine can be understood by the help of biostatistic science. In order to show accurate information for the reader, the most appropriate method of descriptive statistics for displaying data should be chosen.

Measures of central tendency

Measures of central tendency describe the location of the center of a distribution of numerical data. They should bring the most actual and reliable information assuming that this represents the majority of observations on sampling distribution. In clinical research the most used measures of central tendency are mean and median.

Mean (arithmetic average of observations) should be used only when distribution is symmetric (gaussian). This measure is very sensitive to extreme values and therefore few very high values in a non-symmetric distribution could greatly affect the mean resulting in a value which does not reflect the true center of the distribution.

Median reflects the true middle observation in a sample and is less sensitive to extreme values than the mean. The median is not influenced by the non-gaussian distribution of samples and reflects more appropriately the actual central tendency of the sample.

Measures of dispersion

Measures of dispersion describe the spread of numerical data around the central tendency. The most used measures are standard deviation (SD) and percentiles.

SD is a measure of spread of data about their mean. The most important consideration is that using this expressions it is possible to calculate that 67% of the observations lie 1 SD at both sides of mean value and 95% of the observations lie 2 SD at both sides of mean value. Some authors express dispersion as Standard error of the mean (SEM). Two main reasons indicate that this is not appropriate. First, SEM is a function of the sample size (SEM = SD/vn) so it can be made smaller simply by increasing n. Thus the SEM will always be smaller than the SD. Second and more important, the interval (mean ħ 2 SEM) will contain approximately 95% of the means of samples (several samples of observations) but it will never contain 95% of the observations on one sample of individuals. By definition, the SEM pertains to means, not to individuals.

Percentiles are very accurate expression of dispersion because describe precisely the actual percentage of observations below or above each figure. A popular expression is the interquartile range (IQR) that contains the central 50% of observations [1].

In gastroesophageal reflux disease (GERD) patients, reflux of bilio-pancreatic components into the esophagus have been found to be associated with greater mucosal damage, with presence of intestinal metaplasia (Barrett's esophagus) and with the development of esophageal adenocarcinoma.

Bilio-pancreatic reflux into the stomach or the esophagus has been named alkaline refluxate but this has lead to confusion in terminology and definitions.

True alkaline reflux into the esophagus (pH above the neutral value 7) can occur only after total gastrectomy or in patients with achlorhidria due to severe atrophic gastritis or pernicious anemia. On the contrary, subjects without these conditions may suffer reflux of a mixture of acid gastric juice and bilio-pancreatic fluid resulting in a wide range of acidity of the refluxate commonly at pH below 7 and very rarely at pH above this value. In patients with GERD it has been demontrated using simultaneously a pHmeter and an optical ambulatory bilimeter in the distal esophagus, that the acid refluxate (pH lower that 4) can contain bile components up to 20% of time [2]. In this study when pHmeter detected alkaline pH > 7 into the esophagus the presence of bile was found less than 10 percent of times, supporting other observations showing that a pH greater than 7 in the esophagus is mainly due to pooling of saliva and not due to true "alkaline reflux" [3, 4]. Moreover, when the pH is in the normal esophageal range (4-7), bile can be found between 15% and 20% of times suggesting that at this "physiological range" active injuring components might be present.

This should not be surprising since previously Gotley [5] showed similar findings when he studied aspirates of esophageal content in GERD patients, observing that 89% of samples containing bile salts had a pH less than 4. Similarly Collins [6] demonstrated that more than 60% of gastric samples at pH < 3.5 contained bile salts and Stoker [7] showed that most of the esophageal aspirates containing bile salts were at pH < 7.

Therefore the duodeno-gastroesophageal reflux (DGER) can be a frequent finding and occurs almost always at non-alkaline pH (< 7).

Several methods have been used to assess this DGER. Among them the endoscopic visualization of bile into the esophagus and the radiologic observation of DGER by the Capper test (barium instilled into the duodenum through a nasoduodenal tube) are very simple and non-expensive methods but are rather nominal "yes or no" dichotomous data. Since DGER seems to be very frequent in control subjects and in GERD patients, obviously these kind of observations have no accuracy to discriminate normals from diseased subjects.

Hepatobiliary scintigraphic studies using radiolabeled markers (99mTc DISIDA) have been used in studying DGER but there are conflicting results among centers [8-10]. Besides, several shortcomings appear to diminish its reliability: short time of study (6090 min), recumbent position, in-hospital fasting conditions, intermittent nature of bile reflux, patient's movements, overlap of small bowel, stomach and left lobe of liver, etc. It is usually informed as "positive" or "negative" and is very difficult to be expressed in a quantitative scale. Currently it is not longer used in our clinical practice.

Currently two types of tests are used in clinical research to assess DGER:
- measurement of concentration of a selected marker of DGER like bile salts and trypsin in long term esophageal aspirates (it can be expressed in a continuous scale);
- optical ambulatory 24-hour bilirubin assessment (Bilitec™)

Quantitation of duodeno-gastroesophageal reflux by measuring concentration of markers

Bile salts have been used as the most common marker of DGER, however its determination just shows an instant picture of a rather dynamic event. Besides the usual enzymatic method used to assess the presence of this marker has been criticized because it is not specific for bile acid in the postprandial state and therefore should be employed only in fasting periods of study [11]. Table I summarizes the most relevant general findings of several studies of quantitation of bile salts in esophageal aspirates [5, 7, 12-16]. As can be seen there exists a great variability of results among authors. Gotley reported a median of 0 mmol/l which is clearly explained because only 27% of the samples contained bile salts, even when this subgroup showed some extremely high values (˜15,000 µmol/l). Similar situation is observed in the paper of Kauer et al. [15] where the individual data plotted in a graphic clearly shows the majority of observations around the 0 value and a few number of observations reaching high concentration values near 1,000-2,000 µmol/l, which explains the high mean value obtained in this study. Other authors reported greater values mainly during supine period with a narrow dispersion of the data (SEM) around the center (mean). However these authors did not report the proportion of samples negative for bile salts. If we analyze the results following statisticals formulas we can figure out the results: let us assume a sample of 50 observations (n = 50) with a mean concentration = 100 µmol/l ħ 25 µmol/l (SEM). The formula SEM = SD/vn allows us to calculate the SD where:

SD = SEM o vn which is 25 o 7.1 = 177 µmol/l (SD).

This simple operation indicates us the non normal distribution of the observations in this sample, because there is no way to find negative concentration values of bile salts (100-177 = -77 µmol/l) if we substract 1 SD to the mean value.

Table II shows the results of quantitation of bile salts in esophageal aspirates of control subjects and patients with GERD according the severity of endoscopic esophagitis. It can be seen that there is a clear tendency to obtain higher concentration of bile salts in more advanced stages of the epithelial damage. Interestingly the volume of aspiration was not significantly different among groups [14, 16] suggesting that the difference inconcentration of bile salts is not due to a dilution factor. Another very important concern to remark about displaying numerical data can be observed in one report [16] where the median value reported for each group (control, minimal and erosive esophagitis group) is a value different from zero even when less than 50% of the samples contained bile salts. Whether the values reported are just representative of the small proportion of positive bile salts samples it is not clearly stated in the paper.

A third important quantitative data can be drawn from this kind of assessment. If "normality" might be defined from observations in control subjects (i.e.: 95th percentile as upper limit) then the concentration above this figure can be assumed as "abnormally high" (dichotomous data: normal or abnormal). The report of Kauer [15] showed that the proportion of cases with abnormally high concentration of bile salts in esophageal aspirates were 12% in upright period, 25% in postprandial period and 17% in supine period in GERD patients.

Detection of trypsin has been reported [7, 14] as a method for assessment of

Table I. General findings and measures of expression in studies of bile salts in esophageal aspirates.

Table II. Quantification of bile salts in esophageal aspirates in controls and patients with GERD..

DGER. However some limitations make difficult its clinical acceptability as a diagnostic tool. It is a very complicated and expensive test and second, the denaturation of this enzyme at acidic pH is a major drawback for its detection during the majority of reflux episodes.

Quantitation of duodenogastroesophageal reflux by continuous intraluminal detection of markers

Esophageal alkaline exposure pH > 7 during 24-hour pH monitoring was thought to represent DGER. However it has been demonstrated not to be a sensitive and specific indicator of DGER [2-4]. Bile (bilirubin) is the marker for fiberoptic ambulatory assessment (Bilitec™) [17] and has been shown to be a reliable marker of DGER and besides it has a good correlation with the injurious bile-pancreatic components able to harm the esophageal epithelium [18] . The information of this test is displayed similarly to the ambulatory pH assessment, i.e.: frequency of bile reflux during upright, supine or postprandial period, and during total time, number of reflux episodes longer than 5 minutes, the longest episode of bile reflux, percentage of time in which the esophagus was exposed to bile in the same periods and in total time studied. However the percentage of bilirubin detection compared to total time studied is the most commonly used data reported in the clinical research [2, 19-25].

In Table III it can clearly be seen that when median value is used as expression of central tendency, it tends to be much lower than mean value. Again authors using mean value,

Table III. Assessment of bile reflux into the esophagus using Bilitec™ in control subjects and patients with GERD.

express dispersion as SEM instead of standard deviation: this is an obstacle for the reader to get a better idea of the real dispersion. Another consistent observation is related to the increasing amount of bile reflux in more advanced stage of epithelium damage. Patients with Barrett's esophagus and specially those with complicated disease show the greater amount of bile reflux. Csendes [24] found greater bile reflux in short Barrett's compared to controls however the increase was not as high as reported for long Barrett's group in other series, usually above 20%. In this same report bile reflux was similar in controls and in subjects with GERD and cardial intestinal metaplasia. Table IV shows the prevalence of abnormally high bile reflux into the esophagus of patients with GERD compared with control subjects. Again, greater prevalence of excessive esophageal exposure to bile was found in patients with more advanced disease. Similarly, patients with short Barrett's esophagus showed more prevalence of abnormally high bile reflux compared to cases with columnar epithelium without intestinal metaplasia [25]. In summary, quantitative data of DGER can be obtained from current tests available for the clinician. First, part of this information reflects objective behavior of some continuous

Table IV. Prevalence of bile reflux into the esophagus using Bilitec™ in control subjects and patients with GERD.

variables such as concentration of a marker aspirated from the esophageal lumen or percentage of time this marker is detected in the esophagus. Second, if some event is found to present a range of "normality", then the prevalence of patients above this range should provide quantitative discrete information. The display of this data must be objectively shown following basic and appropriate statistical rules, because this will allow the medical community to share valuable information and to get a better understanding of the biological events related to diseases.


1. Dawson-Saunders B, Trapp RG. Basic and clinical biostatistics. Norwalk: Appleton & Lange, 1990.

2. Kauer WK, Burdiles P, Ireland AP, Clark GW, Peters JH, Bremner CG, DeMeester TR. Does duodenal juice reflux into the esophagus of patients with complicated GERD? Evaluation of a fiberoptic sensor for bilirubin. Am J Surg 1995;169:98-104.

3. Singh S, Bradley LA, Richter JE. Determinants of oesophageal "alkaline" pH environment in controls and patients with gastro-oesophageal reflux disease. Gut 1993;34:309-316.

4. Devault KR, Georgeson S, Castell DO. Salivary stimulation mimics esophageal exposure to refluxed duodenal content. Am J Gastroenterol 1993;88:1040-1043.

5. Gotley DC, Morgan AP, Ball D, Owen RW, Cooper MJ. Composition of gastro-oesophageal refluxate. Gut 1991;32:1093-1099.

6. Collins BJ, Crothers G, McFarland RJ, Love HG. Bile acid concentration in the gastric juice of patients with erosive oesophagitis. Gut 1985;26:495-499.

7. Stoker DL, Williams JG, Young RM, et al. Tryptic activity and bile in human oesophageal refluxate. Gullet 1991;1:185189.

8. Matikainen M, Taavitsainen M, Kalima TV. Duodenogastric reflux in patients with heartburn and esophagitis. Scand J Gastroenterol 1981;16:253-255.

9. Warning JP, Legrand J, Chinichian A, et al. Duodenogastric reflux in patients with Barrett's esophagus. Dig Dis Sci 1990;35:759-762.

10. Braghetto I, Gonzalez P, Csendes A, et al. The role of duodenogastric reflux in patients with different degrees of reflux esophagitis. Dis Esophagus 1994;7:99-102.

11. Mittal RK, Reuben A, Whitney JO and McCallum RW. Do bile acids reflux into the esophagus? A study in normal subjects and patients with gastroesophageal reflux disease. Gastroenterology 1987;92:371-375.

12. Smith MR, Buckton GK, Bennett JR. Bile acids levels in stomach and oesophagus of patients with acid gastrooesophageal reflux. Gut 1984;25(5):A556.

13. Johnsson F, Joelsson B, Florén CH, Nilsson A. Bile salts in the esophagus of patients with esophagitis. Scand J Gastroenterol 1988;23:712-716.

14. Stein HJ, Feussner H, Kauer W, DeMeester TR, Siewert RJ. Alkaline gastroesophageal reflux:assessment by ambulatory esophageal aspiration and pH monitoring. Am J Surg 1994;167:163-168.

15. Kauer WKH, Peters JH, DeMeester TR, Feussner H, Ireland AP, Stein HJ, Siewert RJ. Composition and concentration of bile acid reflux into the esophagus of patients with gastroesophageal reflux disease. Surgery 1997;122:874-881.

16. Nehra D, Howell P Pye JK, Beynon J. Assessment of combined bile acid and pH profiles using an automated sampling device in gastro-oesophageal reflux disease. Br J Surg 1998;85:134-137.

17. Bechi P. Fiberoptic measurement of "alkaline"gastro-esophageal reflux:technical aspects and clinical indications. Dis Esophagus 1994;7:131-138.

18. Stipa F, Stein HJ, Feussner H, Kraemer S, Siewert JR. Assessment of non-acid esophageal reflux:comparison between long-term reflux aspiration test and fiberoptic bilirubin monitoring. Dis Esophagus 1997;10:24-28.

19. Champion G, Richter JE, Vaezi MF, Singh S, Alexander R. Duodenogastroesophageal reflux:relationship to pH and importance in Barrett's esophagus. Gastroenterology 1994;107:747-754.

20. Vaezi MF, Richter JE. Synergism of acid and duodenogastroesophageal reflux in complicated Barrett's esophagus. Surgery 1995;117:699-704.

21. Kauer WKH, Peters JH, DeMeester TR, Ireland AP, Bremner CG, Hagen JA. Mixed reflux of gastric and duodenal juices is more harmful to the esophagus than gastric juice alone:the need for surgical therapy re-emphasized. Ann Surg 1995;222:525-533.

22. Fein M, Ireland AP, Ritter MP, Peters JH, Hagen JA, Bremner CG, DeMeester TR. Duodenogastric reflux potentiates the injurious effects of gastroesophageal reflux. J Gastrointest Surg 1997;1:27-33.

23. Burdiles P, Csendes A, Braghetto I and Henriquez A. Manometric characteristics of the lower esophageal sphincter and esophageal exposure to acid and biliary content in patients with GERD and Barrett's esophagus. 6th World Congress of the International Society for Diseases of the Esophagus. Milan, Italy, Aug 23-26, 1995:A163.

24. Csendes A, Burdiles P, Smok G, et al. Clinical and endoscopic findings, and gastric and duodenal reflux assessment in patients with cardial intestinal metaplasia and short Barrett's esophagus compared to controls. Rev Med Chile 1999;127:1321-1328.

25. Csendes A, Smok G, Flores N, et al. Comparison of clinical, endoscopic and functional findings in patients with intestinal metaplasia at the cardia, carditis and short-segment columnar epithelium of the distal esophagus with and without intestinal metaplasia. Dis esophaggus 2000;13:61-68.

Publication date: August 2003 OESO©2015