1424 EDITORIALS GASTROENTEROLOGY Vol. 102, No.

G, DenBesten L, Rubin CE. Endoscopic biopsy can detect DeMeester TR. Barretts esophagus. Comparison of benign
high-grade dysplasia or early adenocarcinoma in Barretts and malignant cases. Ann Surg 1983;198:554-565.
esophagus without grossly recognizable neoplastic lesions. 20. Cameron AJ, Zinsmeister AR, Ballard DJ, Carney JA. Preva-
Gastroenterology 1988;94:81-90. lence of columnar-lined (Barretts) esophagus. Comparison of
9. Haggitt RC, Tryzelaar J, Ellis FH, Colcher H. Adenocarcinoma
population-based clinical and autopsy findings. Gastroenterol-
complicating columnar epithelium-!ined (Barretts) esopha- ogy 1990;99:918-22.
gus. Am J Clin Path01 1978;70:1-5.
21. Nebel TO, Fornes MF, Caste11 DO. Symptomatic gastroesopha-
10. Palley SL, Sampliner RE, Garewal HS. Editorial: management
geal reflux: incidence and precipitating factors. Am J Dig Dis
of high-grade dysplasia in Barretts esophagus. J Clin Gas-
1976;21:953-956.
troenterol 1989;11:369-372.
11. Altorki NK, Sunagawa M, Little AG, Skinner DB. High-grade 22. Thompson WG, Heaton KW. Functional bowel disorders in
dysplasia in the columnar-lined esophagus. Am J Surg apparently healthy people. Gastroenterology 1980;79:283-
1991;161:97-99. 288.
12.Schnell T, Sontag S, Chejfec G, Chintam R, OConnell S, Kur- 23. Tortorello NJ, Earnest DL. Hazards of heartburn; results of a
ucar C. High-grade dysplasia in Barretts esophagus: a report national consumer survey (abstr). Am J Gastroenterol
of experience with 43 patients (abstr). Gastroenterology 1991;86:A1377.
1989;96:A452.
24. Blot WJ, Devesa SS, Kneller RW, Fraumeni JF. Rising inci-
13. Ahnen DJ. Flow cytometric analysis of deoxyribonucleic acid dence of adenocarcinoma of the esophagus and gastric cardia.
content in the gastrointestinal tract. Is aneuploidy more than JAMA 1991;265:1287-1289.
a new ploy? Gastroenterology 1987;93:197-199.
14. Reid BJ, Blount PL, Rubin CE, Levine DS, Haggitt RC, Rabino- 25. Hamilton SR, Smith RRL, Cameron JL. Prevalence and char-
vitch PS. Flow cytometric and histologic progression to malig- acteristics of Barretts esophagus in patients with adenocarci-
nancy in Barretts esophagus: prospective endoscopic surveil- noma of the esophagus or esophagogastric junction. Hum
lance of a cohort. Gastroenterology 1992 (in press). Path01 1988;19:942-948.
15. Fennerty MB, Sampliner RE, Way D, Riddell R, Steinbronn K, 26. Schnell T, Sontag S, Chejfec G. Occurrence of adenocarci-
Garewal HS. Discordance between flow cytometric abnormali- noma in short segments or tongues of Barretts esophagus
ties and dysplasia in Barretts esophagus. Gastroenterology (abstr). Gastroenterology 1989;96:A452.
1989;97:815-820. . Pera M, Cameron AJ, Trastek VF, Carpenter HA, Zinsmeister
16. Ramel S, Reid BJ, Sanchez CA, Blount PL, Levine DS, Neshat AR, Ballard DJ. Esophageal adenocarcinoma: a population-
K, Haggitt RC, Dean PJ, Thor K, Rabinovitch PS. Evaluation of based study of its incidence and the role of Barretts esopha-
~53 protein expression in Barretts esophagus by two-parame- gus (abstr). Gastroenterology 1991;100:A394.
ter flow cytometry. Gastroenterology 1992 (in press).
17. Levine AJ, Momand J, Finlay CA. The ~53 tumour suppressor
gene. Nature 1991;351:453-456.
18. Sanfey H, Hamilton SR, Smith RRL, Cameron JL. Carcinoma Address requests for reprints to: Alan J. Cameron, M.B.,
arising in Barretts esophagus. Surg Gynecol Obstet M.R.C.P., Division of Gastroenterology, Mayo Clinic, Rochester,
1985;161:570-574. Minnesota 55905.
19. Skinner DB, Walther BC, Riddell RH, Schmidt H, Iascone C, 0 1992by the American Gastroenterological Association

Maintaining Intestinal H, Balance: Credit the

A major fermentation reaction of the colonic bacte- nary excretion, and metabolism by other colonic bac-
ria yields about 330 mL of H, per gram of carbohy- teria. In this issue, Christ1 et al. provide the first, reli-
drate metabolized. Because roughly 40 g of carbohy- able, long-term measurements of breath and anal H,
drate normally is thought to enter the colon each excretion in humans.4 Previous attempts to measure
day, enormous (12,000 ml/day) quantities of H, H, elimination per rectum in adult?- have used rec-
should be released into the colonic lumen. Clini- tal tubes that often plug and leak, and generally tend
cally, such H, production has major implications for to be a pain in the rear end, literally for the experi-
flatulence and abdominal distention. From the envi- mental subject and figuratively for the investigator.
ronmental safety standpoint, the anal excretion of The innovative feature of the study by Christ1 et al.
such quantities of H, represents a potential explosion was housing of the subject in a small, airtight room
hazard for both the gas passer and anyone in the gen- equipped with an airlock. Fresh air was passed
eral vicinity.3 Thus, the mechanism by which the through the room at a rate of 100 L/min, and total H,
colon disposes of H, is a problem of more than pass- excretion into the room was determined from the
ing importance. difference between outflow and inflow H, concen-
H, is eliminated from the colon via three routes: trations. Breath H, excretion was independently as-
anal excretion, absorption with subsequent pulmo- sessed by direct measurement of expired air, and fla-
April 1992 EDITORIALS 1425

tus excretion was then calculated from the Three different species of organisms compete for
difference between total and breath H, excretion. H, in the colon: methanogens, which reduce CO, to
While this technique allows long-term measure- methane (CHJl; bacteria that reduce sulfate to sul-
ments of flatus H, in a physiological and comfortable fide; and bacteria that synthesize acetate from CO,
manner, the rapid air flux through the room results and H,. British studies have suggested that sulfate-
in very low H, concentrations (4.0 ppm) during pe- reducing bacteria are the most efficient utilizers of
riods of low H, excretion. Thus, extremely sensitive HZ,13,14but methanogens easily win the competition
and precise methodology is required for accurate de- for H, in the colons of Minnesotans.15 We found that
termination of H, excretion. fecal homogenates of CH,-producing subjects con-
The widespread use of breath H, excretion as a sumed H, about four times more rapidly than did
simple means of assessing H, liberation into the co- homogenates from CH, nonproducers. In addition,
ionic lumen rests on the assumption that a linear incubation of a mixture of both types of feces en-
relationship exists between breath and flatus H, ex- hanced CH, production as the methanogens appar-
cretion. Evidence to support this assumption was de- ently pirated the H, produced by the nonme-
rived largely from a study that used a constant intes- thanogenic feces.
tinal infusion of argon that showed that pulmonary Because 5 mol of gas (4 H, and 1 CO,) is consumed
excretion accounted for 14% of total H, excretionsIn in the process of producing 1 mol of CH,, subjects
the much more physiological study of Christ1 et al. harboring a methanogenic colonic flora should theo-
over a wide range of H, excretion, breath accounts retically have much less flatulence then their non-
for about 50% of total H, excretion. However, at very methanogenic counterparts. In fact, it is surprising
low excretion rates (~50 mL/15 h), most of the H, that subjects who produce CH, excrete any H, be-
was eliminated via the lungs, whereas pulmonary cause in vitro studies have shown that at the high H,
excretion fell to about 15% of the total at very high tensions present in the colon, methanogenic fecal ho-
rates of excretion (>600 mL/15 h). The authors at- mogenates consumed H, at a rate of about 10
tribute this declining importance of breath H, excre- mL - h- - g feces-. Thus, the roughly 500 g of feces
tion to more rapid propulsion of gas through the co- in the colon should be able to consume 5000 mL of H,
lon and/or saturation of H, absorption at high per hour, a value far in excess of any H, excretion
colonic H, concentrations. The former is almost cer- rate ever observed in humans. The apparent explana-
tainly the correct explanation because the passive tion for the only modestly reduced H, excretion ob-
transport of H, should not show saturation kinetics. served in CH,-producing subjects is that fecal homog-
Although this study generally supports the validity enates do not accurately reflect the stirring or the
of breath H, measurement as a simple measure of bacterial composition of colonic contents. Methano-
total H, excretion, it is important to keep in mind gens are commonly present only in the left colon,
that the linear relation between breath and flatus H, whereas H, production occurs throughout the co-
excretion breaks down during periods of very low lon.7*8 Thus, H, liberated in the right colon largely
and very high excretion. escapes the clutches of the methanogens and is avail-
Although elimination of H, via the anus and lungs able for excretion. In addition, the luminal contents
has received considerable attention over the years, it of the right colon may be more efficiently stirred
now appears that utilization of this gas by HZ-con- than was the homogenate. As a result, H, would rap-
suming colonic microorganisms is the major means idly enter the gas space, and the decrease in H, ten-
of disposal of intestinal H,. Evidence to support this sion in colonic contents would markedly reduce the
concept was initially provided by studies showing rate at which bacteria consume this gas. The finding
that 40% of the H, instilled into the isolated colon that lactulose ingestion resulted in far more H, ex-
of conventional rats subsequently could be recov- cretion per gram than did the ingestion of more
ered, whereas nearly complete recovery was ob- slowly fermented, insoluble carbohydrates (e.g.,
served when similar experiments were carried out banana starch)4 presumably reflects the fermenta-
in germ-free rats. Christ1 et al. have now shown that tion of the latter in the left colon, in which poor lu-
the total quantity of H, excreted after ingestion of minal stirring and the presence of methanogens
lactulose was a small fraction of that predicted to be markedly enhanced the efficiency of H, consump-
released by fermentation reactions, and ingestion of tion.
nonabsorbable carbohydrates that are slowly fer- It is apparent that the factors determining a sub-
mented (pectin, banana starch), in turn, yielded only jects rate of H, excretion are far more complex than
a small fraction of the H, liberated by lactulose. the conventional scenario that assumed that the ex-
Thus, H, excretion is not synonymous with H, pro- cretion of this gas is simply a function of the quantity
duction but should be regarded as the absolute pro- of fermentable substrate reaching the colon and the
duction rate minus the consumption rate of H,. numbers of HZ-producing colonic bacteria. In fact, a
1426 EDITORIALS GASTROENTEROLOGY Vol. 102, No. 4

recent study suggested that there are only minor dif- 7. Tomlin J, Lowis C, Read NW. Investigation of normal flatus
ferences in the absolute rate of production of H, by production in healthy volunteers. Gut 1991;32:665-669,
8. Levitt MD. Production and excretion of hydrogen gas in man.
different individuals. Thus, the great individual
N Engl J Med 1969;281:122-127.
variability observed in H, excretion after ingestion of 9. Levitt MD, Berggren T, Hastings J, Bond JH. Hydrogen (H,)
similar doses of nonabsorbable carbohydrates ap- catabolism in the colon of the rat. J Lab Clin Med 1974;84:163-
pears to reflect differences in H, consumption, At 167.
present, therapy for flatulence is limited to dietary 10. Wolin MJ. Fermentation in the rumen and human large intes-
tine. Science 1981;213:1463-1468.
alterations that attempt to reduce the quantity of fer-
11. Gibson GR. A review. Physiology and ecology of sulphate-re-
mentable substrate reaching the colonic bacteria, a ducing bacteria. J Appl Bacterial 1990;69:769-767.
treatment that usually fails because of the wide vari- 12. Lajoie SF, Bank S, Miller TL, Wolin MJ. Acetate production
ety of carbohydrates that are incompletely absorbed from hydrogen and [%]carbon dioxide by the microflora of
by the normal small intestine. Future approaches to human feces. Appl Environ Microbial 1988;54:2723-2727.
13. Gibson GR, Cummings JH, Macfarlane GT. Competition for
this problem may take the form of manipulations de-
hydrogen between sulphate-reducing bacteria and methano-
signed to enhance H, consumption via either reduc- genie bacteria from the human large intestine. J Appl Bacte-
tion of fecal stirring or implantation of Hz-consuming riol 1988;65:241-247.
bacteria into the right colon. 14. Gibson GR, Macfarlane GT, Cummings JH. Occurrence of sul-
phate-reducing bacteria in human faeces and the relationship
ALESSANDRA STROCCHI, M.D. of dissimilatory sulphate reduction to methanogenesis in the
MICHAEL D. LEVITT, M.D. large gut. J Appl Bacterial 1988;65:103-111.
Research Service 15. Strocchi A, Furne JK, Ellis JC, Levitt MD. Competition for hy-
Veterans Administration Medical Center drogen by human faecal bacteria: evidence for the predomi-
Minneapolis, Minnesota nance of methane producing bacteria. Gut 1991;32:1498-
1501.
References 16. Strocchi A, Levitt MD. Factors affecting hydrogen production
and consumption by human fecal flora: The critical role of
1. Weaver GA, Krause JA, Miller TL, Wolin MJ. Constancy of hydrogen tension and methanogenesis. J Clin Invest (in press).
glucose and starch fermentations by two different human fae- 17. Levitt MD, Ingelfinger FJ. Hydrogen and methane production
cal microbial communities. Gut 1989;30:19-25. in man. Ann NY Acad Sci 1965;150:65-81.
2. Stephen AM, Haddad AC, Phillips SF. Passage of carbohy- 18. Flourie BF, Etanchaud F, Florent C, Pellier P, Bouhnik Y,
drate into the colon. Direct measurements in humans. Gastro- Rambaud JC. Comparative study of hydrogen and methane
enterology 1983;85:589-595. production in the human colon using caecal and faecal homog-
3. Sanders T. Man shoots sparks when he passes gas. Weekly enates. Gut 1990;31:684-685.
Word News 199l;October 15:37.
4. Christ1 SU, MurgatroydPR, Gibson GR, Cummings JH. Produc-
tion, metabolism, and excretion of hydrogen in the large in-
testine. Gastroenterology 1992;102:1269-1277. Address requests for reprints to: Michael Levitt, M.D., ACOS
5. Kirk E. The quantity and composition of human colonic fla- Research, Veterans Administration Medical Center, 1 Veterans
tus. Gastroenterology 1949;19:782-794. Drive, Minneapolis, Minnesota 55417.
6. Calloway DH, Murphy EL. The use of expired air to measure This is a U.S. government work. There are no restrictions on its
intestinal gas formation. Ann NY Acad Sci 1968;150:82-95. use.

Community-Acquired Hepatitis C Virus

Hepatitis C virus (HCV), which contains a single- The ability to serologically identify and treat hepa-
stranded RNA encoding a number of proteins, has titis C has raised the consciousness of affected indi-
been identified as the principal cause of non-A, non- viduals concerning the mechanisms by which their
B (NANB) hepatitis in transfusion recipients, hemo- disease may be transmitted to others. Although per-
philics, and injection drug abusers. Its presence can cutaneous or parenteral transmission is readily ac-
be detected by host antibodies to a few of its pro- cepted by these patients as the principal mode of
teins, and by polymerase chain reaction (PCR) as- spread of HCV, their concern is invariably focused
says that make use of specific sequences within the on the potential for contact-associated transmission.
viral genome.3 Patients wish to know whether oral transmission of