Gas/bloat and the physiology of belching

"For the clinician, the bęte noir of the symptoms referable to the digestive tract is 'gas'."(1) Patients can assure themselves of instant rejection by complaining, "Everything I eat turns to gas!" Unfortunately, "gas" and the equivalent symptom of "bloat" is the most common of all gastrointestinal complaints.(2)

Efforts to correlate the "gas" symptom with objective evidence of gas have been futile. Although physicians have " . . . been well indoctrinated that swallowed air is the cause of alimentary tract gas.",(3) they often see patients with flat bellies and gasless radiographs who paradoxically complain, "I feel as though I were going to explode!" On the other hand, a patient may have visible abdominal distention without complaining of gas. Patients with ascites, pneumoperitoneum or extreme obesity seem oddly immune to the complaint of bloating. On the other hand, patients with mechanical bowel obstruction and a genuine gas problem, complain of cramps, not gas.

This puzzle of "gas without gas" has motivated many studies. Although these have yielded exact data on the production, composition, absorption, elimination, and clinical correlations of intestinal gas, they only deepen the mystery. Lasser and associates,(4) for example, using a sophisticated isotope washout technique, found that 12 patients with this complaint averaged 23 ml less intestinal gas (177 ml) than 10 normal controls (200 ml).

In extreme cases, the patient may habitually perform alternate Mueller and Valsalva maneuvers to fill and empty the esophagus with gas (as in esophageal speech) to provoke eructation of gas from the stomach in the belief this will afford relief. In an occasional patient, this can get completely out of hand. The false eructation may become a social liability - an outrageous, gross habit that the patient is apparently unable or unwilling to control.

Most such patients are eventually referred for radiologic examination where, aside from demonstrating of aerophagia(5) if it is extreme, the results have given no clue to the cause of this bizarre performance.

Various theories are (6)offered. Nevertheless, none carry any conviction or suggest an experimental test that would either confirm or refute them. Roth(7) relates belching to aerophagia, splenic flexure syndrome, "magenblase syndrome" and neurosis, considering it only in the context of the excessive belching of aerophagia. He lists aerophagia as caused by most of the disorders of the upper GI tract. He speculates that the habit forms because the ". . . patient experienced some relief of a distress with the eructation of air . . . and thus deliberately induces belching by aerophagia to secure that relief again."

Click for larger image
Click image for larger version

FIGURE VI.1 A-F - Belching produces a HH. Over the course of a lifetime, belching gagging, rapid swallowing, and vomiting generally stretch the PEL beyond its elasic limits - a limit that is also decreaasing with age.
Bockus(8) mentions belching as a frequent symptom of hiatus hernia, but does not describe the actual act. Roth does note that ". . . slightly elevating the chin and extending the neck . . . "(9) is a conscious maneuver to provoke belching, but attributes this act to an effort to swallow air to induce belching.

Earlam(10) lists some quantitative information: the gastric gas bubble is less than 50 ml of air 2-3 ml are swallowed with each bolus; 1,000 swallows/day (70/hr while awake, 7/hr while asleep) would result in total ingestion of 2.0 to 3.5 liters of gas per day. He also reports that the average audible belch contains 20-80 ml and requires simultaneous contraction of the abdominal musculature and relaxation of the sphincter.

One need not consult standard references to become acquainted with the mechanics of belching. Self-observation will go a long way. Thus one finds he unconsciously leans forward, head up, neck extended. The larynx ascends as in swallowing and gas is released. The forward inclination of the thorax is particularly noticeable if one is restrained by the shoulder strap of a seat-belt.(11) Unfortunately, this introspection gives no clue to what is going on at the lower end of the esophagus. It would be much better to study the process fluoroscopically.

Inducing a belch at fluoroscopy

Chance fluoroscopic observation of a belch initially aroused my interest in the problem. Thereafter, I was frustrated by the infrequency with which I could study the process further. Chance observations were infrequent and when they did occur, one seldom had the area of interest in the field of view, good wall coating, cine camera on, etc. One cannot just wait for a belch to happen - fluoro time will run out with the examination uncompleted.

Click for larger image
Click image for larger version

Belch: LMC jerks the stomach into the chest to the end of the tethering PEL
Eventually, as the Cannon-Dougherty (CD) reflex mechanism became clear, I realized that, because the CD receptor was on the posterior wall of the stomach, simply rolling the patient over into the prone (RAO) position after the dC maneuver would remove the inhibitory effect that submersion of the receptor for this reflex has on sphincter opening. This had the desired effect, eliciting a belch in 20-30% of patients.

LMC induction of eructation

There appear to be two factors involved in the success of this maneuver, neither of them the amount of air in the stomach:

1. The gas or bloat symptom. Such patients yield the bulk of the positive responses.

2. The amount of water employed. Using 180-250 cc of water and rocking the patient back and forth to slosh it about the fundus seems to increase the yield.

With this technique, there were many opportunities not only to secure sequence spot and cine films but also to ask patients whether the sensation experienced during longitudinal muscle contraction (LMC) was the same as the sensation they were calling "gas."

The first visible warning of an impending belch is a slight conical tenting of the otherwise hemispherical outline of the fundus or the formation of a hooded appearance of the fundic mucosal folds.(12) This may vanish or become more pronounced. If the latter, a hiatus hernia, if present, becomes drawn upward to the full length of the tethering phrenoesophageal ligament. The whorls of mucosal folds in the fundus then straighten and point to the gastroesophageal junction and that junction assumes the shape of a trumpet bell or, more accurately, an alpenhorn. The diaphragm is tented and, as a result, loses its sharp outline because the central ray of the beam passes through the tent rather than grazing the diaphragmatic dome.

This bell shape is the shape of an elastic membrane under traction. The greater the force applied to the center of the membrane, the more acute the apex angle of the cone of tensed membrane. The perihiatal region of the diaphragm itself may also be tented upward. These changes are manifestations of a LM contraction that, in extreme cases, may shorten the esophagus 36% or more.

The sphincter then opens and stomach gas instantly inflates the esophagus. If the voluntarily controlled superior constrictor opens, gas escapes into the pharynx, is eructated and the fundus collapses. If not, an en masse contraction of the circular muscle and/or a peristaltic wave forces gas back into the stomach.

In this way, a striking event in patients who belched under fluoroscopic study was a preparatory strong contraction of the LM (LMC). Questioning the patients as it occurred established that the gas/bloat sensation was simultaneous with the signs of LMC.

This portion of the act of belching, unlike gagging or vomiting, is quite deliberate and usually occupies several seconds. LMC may be sustained for a considerable time if, for any reason, the superior sphincter does not open. As much as 8-10 seconds is not unusual and in an exceptional instance it may be sustained 30 seconds or longer allowing ample time to question the patient.

Sphincter release does not follow every preparatory LMC. The LMC may simply subside or there may be a partial subsidence followed by contraction leading to an almost rhythmic ebb and flow of tension on the GE junction marked by varying shape of the trumpet bell. The greater the force applied to the center of the membrane, the more acute the angle of the apex of the cone formed by the stretched membrane.

Click for larger image
Click image for larger version

On cine films in such cases, one can see the sphincter margins alternately approaching and receding from each other as though prepared for an instant closure whenever some very delicate balance of forces is destabilized. This process is a clear and convincing demonstration that the LM opens the sphincter: the shorter the LM, the higher the GE junction and the more patent the sphincter.

Gas/bloat is a misinterpretation of LMC

Sometimes, the conscious control of belching can be dramatic as in the following case:

11/4/66 H.D. 45151/M-426 Fluoroscopic note: There was considerable aerophagia and a large amount of gas accumulated in the stomach. A "hiatus hernia" was present that was ideally provoked by having the patient belch, a feat at which he was unusually proficient and could perform on command. This produced shortening and elevation of the esophagus, evoked the HH and was associated with a descent of the diaphragm resulting in a violent jerking motion of the stomach back and forth through the hiatus. Cine films confirmed and showed free cardio-esophageal reflux.

The tension noted during LMC is perceived at the conscious level as gas although the patient's description of the sensation may be quite variable. Most had trouble defining the sensation, but such expressions as "It feels full.", "Kind of pulling.", "Like a pressure.", or "Bloated." were used. When the cardia tents, if a leading question, "Does it feel as though you were going to belch?" is asked, the response is usually an unequivocal "Yes." From there it is a small step to the conclusion that the sensation preceding a belch is what patients mean by "gas" and similar vague fornulations.

Patients often expressed surprise that I knew they were about to belch or that I could see a belch although they had eructated silently and politely! The uniformly positive identification of LMC with the gas sensation left no doubt about the cause of this mysterious symptom. The identification was convincing after relatively few cases because the patient could identify the symptom with the event as it happened. That is, he did not have the sensation throughout the examination, but at the exact time I observed traction on the gastric fundus.

The LM tension has many interpretations other than gas and bloat as in the following case.

LM051846: This 50 year old male truck driver complained almost constantly of the sensation of a mass "about the size of a large potato" beneath the right diaphragm. Occasional heartburn. Ultrasonography revealed a normal gallbladder and minimal evidence of fatty infiltration of liver. Upper GI demonstrated grade ii reflux (asymptomatic), grade ii duodenitis and antral gastritis. When the fundus tented after the dC maneuver he was surprised I was able to tell exactly when he was having the mass sensation.

He was given a 10 mgm capsule of Nifedipine with instructions to chew and swallow it the next time the symptom was severe. He reported back within the hour, "I didn't take it on the drive home because I wasn't sure what it would do to me, but I had complete relief of the symptom within minutes of swallowing the pill. I feel completely relaxed."

Gas, or better, LM tension is a symptom because there is a condition in which LM tension is hyper - "longitonia" if you will. The LM contracts, increasing its tension on the diaphragm, even when there is no physiologic need for it to do so. It does not require the stimulus of gas distention of the stomach to exert traction on the PEL. A degree of traction great enough to force the sphincter by vector resolution will also tension the diaphragm. For this reason, patients with reflux usually also complain of gas/bloat.

The reason patients misinterpret LM tension on the diaphragm is that LM contraction is an event that frequently precedes eructation of gas. It is independent of whether or not there is actually gas in the stomach. This is why the gas sensation does not correlate with how much gas in the stomach. It is purely a function of LM tension.

As was noted earlier, extension of the neck is a belch-facilitating maneuver. This is true because extending the neck applies tension to the esophagus and may also produce a stretch reflex.

From the point of view of patient care, I believe the identification of the "gas" symptom with LM tension is of value. At a minimum, efforts to decrease gas formation or to adsorb gas on activated charcoal (It has been done!) are futile. Pharmacologic efforts to reduce LM tone would be more rational.

Click for larger image
Click image for larger version

Gas-bloat with LMT. Male 56, pyrosis, bloat, nocturnal laryngospasm, lost teeth at 26. An uneffaced sphincter serves for measurement of LMC. Note shortening from A to B. The esophagus was constantly shortened. This causes traction on the diaphagm producing the “gas-bloat” symptom. This patient also has an enalarged lingual tonsil, gr. 2 vallecular sign and the p-wave was ineffectual.
It seems paradoxical that ingestion of bicarbonate of soda - a substance that will generate gas when it contacts gastric HCl - often relieves the sensation of gas. It must do so by lessening the degree of gastric irritation, suggesting that the latter, whether it be by excessive acid, carminative, toxic substance, etc., may be a stimulus to LMC.

There seems at least a lay consensus that certain foods produce gas. Seed catalogs advertize burpless varieties of cucumbers. Is it possible that such foods contain an active ingredient that heightens LM tone?

Belching and mass contraction of the longitudinal muscle

Like the LMC of nausea, pyrosis and vomiting, there is no peristalsis associated with the LMC of belching except for a post-belch cleanup wave. The force of LMC can be gauged by the size of the conical tent; the higher and thinner the tent, the greater the traction that is being applied to it.

It is noteworthy that, although in peristalsis LM and CM contraction are precisely integrated, in en masse contraction they are can be independent. In pathologic circumstances - particularly diffuse esophageal spasm - simultaneous LM and CM en masse contractions also can occur. In all of these cases, it is striking how the esophagus, to play its many roles, coordinates its several functional elements in sharply different ways.

Because it enables us to correlate pressure relationships in the esophagus with the events just described, it is useful to review the phenomena McNally, Kelly and Ingelfinger(13) recorded from the 2-dimensional viewpoint of the manometer. They found that insufflation of the stomach with air via catheter raised the intragastric pressure to 5-7 mm Hg. Within the 200-1600 cc range, intragastric pressure was independent of how much air was introduced, suggesting that distention rather than pressure is a stimulus to belching.

Manometrically, the escape of air from the stomach into the esophagus was signaled by a sudden equalization of gastric and esophageal pressures [i.e., sphincter release]. In cases where it was possible to record the intrasphincteric pressure, it was found that it could be maintained at values equal to the intragastric pressure for periods of 14-110 seconds before the escape of gastric gas into the esophagus. In two cases ". . . reflux occurred 88 and 96 seconds respectively after intragastric pressures had exceeded those recorded in the sphincter zone." [Emphasis added, i.e., pressure gradients were not sufficient to open sphincter.] The authors were puzzled that ". . . . no. . . . correlation between increased gastroesophageal pressure gradients and the appearance of simple reflux was noted."

In a later study, Sigmund and McNally(14) found a gradual reduction in sphincter pressure preceding a belch induced by essence of peppermint, a carminativet. Perceptively, they drew a significant conclusion: "The gradual decrease in intrasphincteric pressure suggests an active relaxation of the sphincter rather than a passive one, secondary to the increased gastric pressure forcefully distending the sphincter." [Emphasis added.] That active agent, however, is the LM not intragastric pressure.

Both radiologic and manometric techniques demonstrate that release of gas via the superior constrictor may be delayed for many seconds or inhibited entirely. Kahrilas et al.(15) using manometric techniques found that the upper esophageal sphincter (UES) released gas 1-10 seconds after the common cavity effect [gastric pressure = esophageal pressure] that denotes release of the LES. The reflex controlling its relaxation could distinguish between refluxing fluid and gas, probably by sensing the spatial and temporal characteristics of release of gas into the body of the esophagus. As it was unimpaired by mucosal anesthesia these authors were able to exclude a mucosal pH receptor.

Because the sphincter is closed in these preliminary stages of belching, the state of the circular muscle of the body of the esophagus in not seen directly. However, the instant the sphincter opens, the body of the esophagus balloons to its full diameter. From this, we can be certain that the circular muscle behind the sphincter is relaxed at the time of sphincter release. This establishes that LM contraction is not accompanied by CM contraction in belching.

The esophageal lumen collapses as the gas leaves via the superior constrictor. Fluoroscopically, I find it impossible to decide whether this collapse is due to an en mass contraction of the circular muscle or to a momentary Valsalva maneuver that occurs simultaneously with the belch. The latter could force gas from the esophagus even without circular muscle contraction. There is manometric evidence that, like the LM, the circular muscle is not restricted to one mode of contraction. McNally et al. interpreted a simultaneous spike in the tracing from a catheter placed in the colon as an indication that a brief Valsalva maneuver was occurring. This, however, is subject to another interpretation as will be seen in the next chapter.

Belching and the CD receptor

We have seen that a sustained forceful LM contraction inetiates sphincter opening and allows the escape of gas from the stomach. What terminates the egress of gastric contents? Several lines of evidence indicate that then acid/pepsin contacts a fundic recptor LM contraction is inhibited and the sphincter, released from its dilating force, closes.

Often, a belch is not a single event. It may be one of a sequences of eructations of which the first is usually the largest (noisiest) followed by cleanup secondary or tertiary burps. This is particularly the case in infants. It seems that if they are going to spit up feeding, it occurs on the secondary burp.

Click for larger image
Click image for larger version

Belch: (A) Before, (B) after. Although the release of air from the stomach may be explosive, unless the superior constrictor releases, the esophagus may remain air-distended for 8-24 seconds, affording an excellent opportunity to note the traction effects of LMC and their correlation wih the opening of the sphincter. Barium is dilute from he dC test which has turned off the CD receptor that acts as a guardian of the gate to prevent reflx when submerged in acid/pepsin. The trumpet shape explains the orad directed wedge shape of Mallory-Weiss tears.
Rarely, I have had an opportunity to view a belch in the upright position rather than in the RAO (supine) position in which belching is usually induced. This provided an excellent opportunity to see the exquisite timing of these events. As the fluid level in the fundus of the stomach rose toward the sphincter with escape of the gas, the sphincter snapped shut just in time to prevent the escape of gastric contents. Viewing this left little room for doubt that the sphincter knew the fluid level was coming at it. When a large amount of gas escapes, gastric contents may splash the periesophageal area and end the eructation prematurely. A correcting burp soon follows.

These observations only confirm what everyone knows from personal observation - that normally, even when the stomach is full after a meal, gas can be released from the stomach very forcefully without regurgitating fluids. Parents burping infants discover this fact several times a day. Mechanistically, the problem is to learn how the sphincter (or better, the LM that is holding it open) knows that in a few milliseconds the esophagus is going to be doused with gastric contents and finds out in time to do something about it.

The experiments of Dougherty et al.(16) demonstrated the elaborate arrangements ruminants have evolved to prevent reflux of gastric fluids while allowing free escape of gas. The reticuloruminal fold acts as a dam to prevent fluid from reaching the esophageal orifice. Gas, of course, can travel over the dam to exit the stomach. Less elaborate, but nonetheless effective measures perform the same function in man.

By the constraints placed upon it, the receptor for this reflex closure should be a chemoreceptor. Yet, if one postulates such a sensor, the rapidity of its transmission to the motor arm of the reflex is difficult to square with the time it takes for a chemical to diffuse to the depth of the sensor and through any mucus coating it. On the other hand, gastric irritation from certain foods may cause increased mucus production that, by coating the receptor, puts the watchdog to sleep with resulting reflux and heartburn. Such a mechanism would account for the latent period between a dietary indiscretion and the onset of heartburn.

The alternative to a chemoreceptor is some system of sensors that continually monitor the shape of the stomach - a possibility, perhaps, as the shape of a stomach partially filled with air would, for the same volume of contents, be different from one filled entirely with fluid.

There are, however, objections to the latter hypothesis. For one thing, the analogy to the CD receptor is too strong. It will be recalled that the nerve network demonstrated by Dougherty and his co-workers is near the esophageal orifice. This is more in keeping with a chemoreceptor activated when contacted by gastric contents than it is with shape sensing. The latter would require a diffusely distributed net of sensors not localized to this specific region at all.

Surgical causes of the gas/bloat symptom

The gas/bloat symptom is inseparable from the act of belching because the LM tension that produces sphincter release of necessity also exerts traction upon the diaphragm. The symptom is produced whether or not sphincter release is achieved. We have seen that the association of the sensation of diaphragmatic tension with belching causes misinterpretation of this tension with its normal result - an eructation of gas.

Without really intending to do so, surgeons will often perform experiments for us on a scale so vast they have great statistical significance. One way of surgically producing tension on the diaphragm is to pull the GE junction down below the diaphragm and suture the stomach around it so it will not retract to its normal position. By thus "taking a tuck" in the esophagus, a more or less continuous pull on the diaphragm is created as the hypertonic LM tries to pull the plicated fundus through the hiatus. Thus, the gas bloat syndrome which may persist for many years after a Nissen fundoplication. An Angelchik prosthesis often produces the same effect(17) for the same reason.

Infantile colic

Before leaving the subject, there is one aspect of belching that probably deserves discussion - infantile colic. The colicky baby screams, apparently for no reason, as the parent walks the floor wondering what to do to relieve the child's evident misery. The screams seem senseless and different in kind from those of a hungry baby. The infant may be soothed by again feeding it but soon it is having another attack of colic.

Careful mothers learn never to put the baby back in its crib before it has been burped, but this is not always a solution. It is a trial to the parent because the burp may not come until it is nearly time for the next feeding.

Perhaps because of its association with difficulty in burping, the implicit assumption is that colic is due to the excessive air that, failing a satisfactory burp, must be passed through the GI tract. However, radiologists, who see many infant chest and abdomen radiographs, know that non-colicy infants normally have quantities of gas in the small and large bowel. If gas leaves the stomach, there is nothing to stop it being eliminated as flatus. As adults, we are aware that it is no problem to pass any required amount of gas and that it is not painful to do so. Why should it be any different for infants?

When we think of a bowel cramp, we think of a sustained contraction of the circular muscle, because all of the bowel cramps with which we are familiar are circular muscle cramps. Circular muscle spasm can be seen with the fluoroscope, e.g., the cramping pain often associated with barium enemas in a patient with a spastic sigmoid, the pylorospasm noted with delayed gastric emptying - or felt by the examiner as with the broom handle descending colon palpated in patients with spastic colons. We are just not accustomed to seeing LM spasm in the bowel, much less to palpating it.

This produces a typical clinical impasse: a set of symptoms leads the physician to expect a corresponding set of objective findings. The expected findings do not appear. Conclusion: "No disease; over-anxious mother."

Yet colic is too common, too real and too disruptive to the lives of young parents to be dismissed in this fashion. The same parents may alternate children with and without colic. To me it seems far more probable, because of its association with burping problems, that infantile colic may be the simply the pediatric equivalent of bloat. Unfortunately, infants cannot describe their symptoms, but it would add weight to this supposition if the symptom responded to medication as did my patient's "gas" symptoms.


The mechanism at the lower esophagus must be able to permit eructation of air while denying egress to gastric fluid contents. Two elements are involved:

1.) The CD reflex, which inhibits LMC when activated, must be turned off.

2.) The sphincter must be opened by vector resolution of the force of LMC. These two mechanisms account for the perfection of control of this physiologic function.

"Gas," a symptom that is the bane of the clinician, is due to LMC. LMC applies tension to the diaphragm via the PEL. Because the resulting sensation normally precedes a belch, it is identified at the conscious level as a sensation of gas, bloating or epigastric fullness. It can be seen radiologically as a tenting of the cardia, especially if a modification of the usual deCarvalho test technique is used to induce belching by turning off the receptor for the CD reflex.

Postoperative gas/bloat is due to surgically tensioning the esophagus by fundoplication or insertion of a prosthesis. The problem of the colicky baby is discussed in the light of the relation of LMC to belching. It seems that "baby bloat" would explain the phenomena better than current rationales.

As would be anticipated, patients who have a slack, elongated esophagus as in achalasia are unable to belch(18) Vagal cooling abolishes transient sphincter relaxation in dogs by abolishing LMC. The deliberate eructations I have described in this chapter, although most easily studied fluoroscopically are not the most common. There is a second method of sphincter release, alone or in cooperation with LMC. This will be discussed in more detail in the following chapter.


Last Updated July 27, 2007 by David PJ Stiennon

1. 1. Roth, J.L.A In: Bockus Gastroenterology, Fourth Edition, Vol. 2, Ed. Berk, J.E., WB Saunders Company, Philadelphia, 1985.

2. 2. Lasser, R.B., Bond, J.H. and Levitt, MD, The role of intestinal gas in functional abdominal pain. N Engl J Med 1975; 293:524-6.

3. 3. Anon., Gas. Editorial, N Engl J Med 1969; 281:164-5.

4. 4. Lasser, et al., op cit..

5. A misnomer. Little if any of the air aspirated into the esophagus reaches the stomach.

6. 5. Bockus, H.L., Gastroenterology, 2nd Ed., W.B. Saunders, Philadelphia, 1963.

7. 6. Roth, J.L.A. In Bockus, H.L., Gastroenterology, 2nd Ed., W.B. Saunders, Philadelphia, 1963.

8. 7. Bockus, H.L., op cit.

9. A maneuver that stretches the esophagus.

10. 8. Earlam, Richard, Clinical tests of oesophageal function, Grune & Stratton, New York, 1975.

11. Perhaps this is one factor in citizen resistance to wearing them!

12. This hood is the mysterious "image en burnous" of Jutras, et al.

13. 9. McNally, Edmund F., Kelly, J. Edward and Ingelfinger, Franz J., Gastroenterology 46: 245-9, 1964.

14. 10. Sigmund, Charles J. and McNally, Edmund F., The action of a carminative on the lower esophageal sphincter. Gastroenterology 56:13-18, 1969.

15. 11. Kahrilas, P.J., Dodds, W.J., Dent, J.B., Wyman, J.B., Hogan, W.J. and Arndorfer, R.C., Upper esophageal sphincter function during belching. Gastroenterology 91:133-40, 1986.

16. 12. See the chapter on the Cannon-Dougherty reflux for references

17. 13.Burhenne et al., op cit.

18. 14. Holloway, R.H., Dent, J. and Wyman, J.B., Impairment of belch reflex in achalasia: evidence for neural mediation of transient lower esophageal sphincter relaxation. (Abstract) Gastroenterology 91:1055, 1986.