| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Original Research |
1 Section of Pulmonary, Critical Care, and Sleep Medicine, Department of
Medicine, M/C 719, University of Illinois at Chicago, 840 S Wood St., Chicago,
IL 60612-7323.
2 Present address: Division of Pulmonary Medicine, Rochester General Hospital,
Rochester, NY.
3 Section of Cardiology, Department of Medicine, University of Illinois at
Chicago, Chicago, IL.
Received November 21, 2007;
accepted after revision April 4, 2008.
Address correspondence to D. E. Schraufnagel
(schrauf{at}uic.edu).
Abstract
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
SUBJECTS AND METHODS. The maximal air-containing esophageal lumina on each section of standard CT scans of 110 subjects were measured. These subjects came from a cohort of 10,132 self-reported healthy individuals who underwent CT for measurement of coronary artery calcium. Measurements were interpolated to account for variation in the length of the thoraces.
RESULTS. Greater than 60% of the esophageal segments contained no air. On average the maximum air column was 10.5 (SD, 5.0) mm. Only 7.9% of the lumina were > 10 mm. Only 2% were > 15 mm, and only 0.2% were > 20 mm. The average number of lumina > 10 mm as a proportion of the entire esophageal length was 8% (14%). The average size at the carina was 2.6 (4.1) mm. In the upper 30% and from 61% to 75% down the length of the esophagus, < 5% of the lumina were > 10 mm. Less than 3% of subjects had air in the lowest two sections, indicating that the normal lower esophageal sphincter was closed.
CONCLUSION. Esophageal air of > 10 mm should be considered abnormal in all segments except between the cardiac ventricles and lower esophageal sphincter. In this area, > 15 mm should be considered abnormal. An air–fluid level is abnormal.
Keywords: chest CT • esophagus • normal • radiographic anatomy
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
With CT, Goldwin and colleagues [3] found 64% of normal esophagi contained air. Others have commented that air in the esophagus is a normal finding [4] and that an air column greater than 3 cm indicates esophageal hypotonia [2], but an in-depth analysis has not been done. Bhalla and colleagues [5] used a definition of esophageal dilatation as an air-filled lumen more than 10 mm in the coronal plane. We also found no mention in the literature of the frequency of finding food or fluid in the esophagus.
We undertook this study to define the CT appearance of the normal esophagus or, more precisely, the normal esophageal air column. In this case, "normal" means a true population standard as part of a normal distribution curve, not necessarily free of abnormality. We wanted to determine the frequency, extent, and distribution of air in the esophagus in healthy subjects. Once the normal distribution is established, we can use this to diagnose disease and study whether the esophageal size is increased in lung disease that may be associated with aspiration.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
CT and Radiologic Measurements
Images were obtained at full inspiration on a C-100 scanner (Imatron, GE
Healthcare) using 1- to 2-mm collimation. They were reconstructed to a
512-pixel matrix using a sharp kernel and a 26-cm display field of view. They
were transferred to an image analysis workstation (Netra, ScImage). Only
standard mediastinal windows (width, 396 HU; level, 44 HU) and lung windows
(width, 1,465 HU; level, –498 HU) were viewed. The study was restricted
to the esophagus after it entered the thorax. The image was set to lung
windows and stepped downward until one full section into the lung, which was
an area free of influence of the extrapulmonary apical soft tissue. All
sections were 6 mm.
The view was then changed to the mediastinal window. The largest linear measurement of the esophageal air column of each transverse section was measured (Fig. 1) until the esophagus entered the stomach, which was usually at the first abdominal CT section. If no air column was visible, it was recorded as zero. If the esophagus was folded and had two air densities, only the largest was measured because the measurement line was restricted to air only. If the esophagus ran parallel to the plane of view, the measurement would be made perpendicular to the long axis of the esophagus. The internal measurements were recorded for each section. In addition, the esophageal lumen at a constant point, the carina, was recorded for each subject. The carina was determined on the lung window view when it first appeared, but the measurements were made on the mediastinal window. The lower esophageal sphincter was recorded as closed if the lumen was zero (no air) in the last section before entrance into the stomach. A hiatal hernia was recorded as present according to standard criteria: The stomach was identified as present above the diaphragm by gastric rugae, larger lumen, less tubular uniformity, and convergence of the gastric folds above and below the diaphragm [2, 10]. Food was scored as present if heterogeneous material was found at any point in the esophagus. A fluid level was tabulated if there was an air–fluid level at any place in the esophagus.
|
Data Analysis and Interpolation
The luminal measurements were recorded as the exact distance from the first
full thoracic sec tion and also as a percentage of total thoracic esophageal
length to account for the differences in body height. The longest
intrathoracic esoph agus was 26.4 cm (44 sections). Subjects with shorter
thoraces had fewer measurements and therefore "missing values"
compared with the longest. To compare the esophagus at the same level for each
subject, the missing values were interpolated at regular intervals in
proportion to the total distance. The maximum number of measurements was 44. A
thorax with only 40 esophageal sections would have four missing measurements
compared with the longest. These missing values were evenly spaced at sections
number 11, 22, 33, and 44. Each inserted missing value moved the other measure
ments distally to fill out the 44 data points. Each section (100% divided by
44) represented 2.27% of the distance down the thoracic esophagus and allowed
infer ences about the esophagus without regard to body length.
Comparisons between men and women were performed with the Mann-Whitney test. Spearman's correlation was performed to see if age was related to luminal size, and the esophageal measurements were correlated with each other. All tests were performed on SPSS software, version 15.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Of the 3,983 sections in which the esophageal lumen was measured, the maximum air column was 23.2 mm. The average maximum for a given subject was 10.5 (5.0) mm. The average lumen at the carina was 2.6 (4.1) mm. The carina occurred at 38.5% (1.9%) of the length of the esophagus in the thorax, which was at about section 17. The average lumen of all segments of all subjects was 2.7 mm (SD, 2.3 mm), although this number may be misleading because 60% of the esophageal sections contained no air. (The median lumen was 0.) Figure 2A, 2B, 2C, 2D, 2E shows the CT images of representative sections down the thorax to help relate esophageal percent distance to other landmarks.
|
|
|
|
|
|
|
|
|
Removing the sections containing no air (scored zero) showed that much of the variation in the different levels of the esophagus was related to an increase in the number of closed segments. However, this did not account for the two "bulge areas" above the lower esophageal sphincter and behind the upper part of the heart. Only 10% of the lower esophageal sphincter areas contained any air.
In only one subject were all esophageal segments closed. Two individuals were outliers. The measurement of all segments of one subject was 4 SDs from the mean. The esophageal measurements were highly correlated with measurement of the neighboring segments, as one might expect, but were poorly correlated with distal segments. For example, section 1 was only significantly correlated with sections 1–7. Section 2 was only correlated with sections 1–9. Sections 3 and 4 were only correlated with sections 1–13, and section 44 was only correlated with sections 41–43. The same was generally true with other sections but with more variation in the middle. There was no correlation with age and any esophageal lumen.
An air–fluid level was found in 1% and food was found in 9% of the subjects. The lower esophageal sphincter was open in 3%. A hiatal hernia was present in 6% and the average distance of the hiatal hernia was 2.8 (SD 9.9) mm. None of these was different between the sexes.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
We found that the criterion for dilatation of Bhalla and colleagues [5] of 10 mm is reasonable in the first 30% of the esophagus because less than 5% of persons have lumina that large. The same is true for the segments from 61% to 75% of the total length. Less than 3% of subjects had air in the lowest two sections, indicating that the normal lower esophageal sphincter is closed. In the area above the esophageal penetration of the diaphragm and below the heart, however, dilatation of greater than 10 mm occurred in greater than 15% of subjects. The most dilated segments, from 89% to 93% down (sections 39–41), had 6.5–9.5% of their segments greater than 15 mm, although only 1% of segements (one section, number 40, 91%) had measurements greater than 20 mm. From this, we suggest that lumina up to 15 mm should be considered normal in the area below the ventricles and above the diaphragm, but any measurement of 20 mm or more should be considered abnormal.
The presence of two outliers would be expected in a screened population and reflects that 1–2% of normal subjects have esophageal air column enlargement. If the outliers are removed, the same conclusions are reached, but the strength is slightly increased.
The sex differences are interesting in that women have more closed segments posterior to the heart than men. However, men have larger hearts than women [11], which does not support the concept that the esophageal segments are closed behind the heart because of the pressure of the heart on the esophagus.
There were several potential problems is developing these standards. First, thoracic length, and therefore esophageal length, varies; and this variation obscures meaningful interpretation of the different esophageal regions unless accounted for by our interpolation methods. Second, are the air column differences due to lumen size differences or differences in number of closed segments? Accounting for the closed segments reduces much but not all of the variation in the segments.
This study defines the CT appearance of the normal esophagus as part of a normal distribution curve, which includes abnormalities at the frequencies that they occur in the general population. These abnormalities are then screened out by calling measurements within 2 SDs of the mean as normal. This is comparable to establishing normality by measuring a blood laboratory value of a random sample of healthy persons and taking those values that fall within 2 SDs of the mean as normal.
CT was performed in these subjects for an entirely different purpose, but we believe this does not detract from the study. Actually, we see this as freeing the data from bias because we had no preconceived notion of what the esophageal measurements would be, and we had no influence on the subject selection. This would be better than, for example, a prospective series from a gastroenterology clinic that would be biased by referral or a study prospectively recruiting volunteers for esophageal CT that would bias the study because of the subjects' symptoms or concerns about their esophagi. We made no assumption that the subjects had an esophagus free of disease but sought a cross-section of healthy Americans. On the limited parameters measured, our study population compared favorably to the NHANES population.
For recommendations of normality, we can say that air is normal in the esophagus, but distention of more than 10 mm at a prespecified fixed point, such as the carina, is uncommon. The 2-SD upper limit of normal for all subjects at the carina was about 11 mm (mean, 2.6 mm; SD, 4.1 mm). If the two outliers are removed, the average at the carina was 2.6 mm (4.0 mm), moving the 2-SD criterion for normal only slightly over 10 mm. Any segment over 20 mm should be considered abnormal. The normal lower esophageal sphincter should be closed. In the upper 30% of the esophagus (to about the top of the aortic arch) and from 61% to 75% of the length of the esophagus (behind the base of the heart) less than 5% of the lumina were greater than 10 mm, which could also be a standard.
The greatest luminal dimensions and variation occur in the area between where the diaphragm first appears and where the esophagus enters the abdomen. Here, an upper limit of normal should probably be considered 15 mm, although 10% of these healthy subjects had at least one segment that exceeded 15 mm in this area. An air–fluid level on any section should be considered abnormal but food should not.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
