Article Text

Download PDFPDF
Original article
Hiatus hernia in healthy volunteers is associated with intrasphincteric reflux and cardiac mucosal lengthening without traditional reflux
  1. Elaine V Robertson1,
  2. Mohammad H Derakhshan1,
  3. Angela A Wirz1,
  4. David R Mitchell1,
  5. James J Going2,
  6. Andrew W Kelman1,
  7. Stuart A Ballantyne3,
  8. Kenneth E L McColl1
  1. 1 Section of Gastroenterology, ICAMS, University of Glasgow, Glasgow, UK
  2. 2 University Department of Pathology, University of Glasgow, Glasgow, UK
  3. 3 Department of Radiology, Queen Elizabeth University Hospital, Glasgow, UK
  1. Correspondence to Professor Kenneth E L McColl, BHF Glasgow Cardiovascular Research Centre, University of Glasgow, 126 University Place, Glasgow G12 8TA, UK; Kenneth.McColl{at}glasgow.ac.uk

Abstract

Background and aims Hiatus hernia (HH) is a key mediator of gastro-oesophageal reflux disease but little is known about its significance in the general population. We studied the structure and function of the gastro-oesophageal junction in healthy volunteers with and without HH.

Methods We compared 15 volunteers with HH, detected by endoscopy or MRI scan, but without gastro-oesophageal reflux disease with 15 controls matched for age, gender and body weight. Jumbo biopsies were taken across the squamocolumnar junction (SCJ). High-resolution pH metry (12 sensors) and manometry (36 sensors) were performed upright and supine, before and after a meal. The SCJ was marked with an endoscopically placed clip and visualised fluoroscopically.

Results Cardiac mucosa was longer in volunteers with HH (3.5 vs 2.5 mm, p=0.01). There was no excessive acid reflux 5 cm above the upper border of the lower oesophageal sphincter (LOS) in either group but those with HH had short segment reflux 11 mm above the pH transition point after the meal when supine (pH<4 for 5.5% vs 0.3% of time, p=0.01). The SCJ and pH transition point were proximally displaced within the gastro-oesophageal junction in those with HH versus controls (p<0.05). The pH transition point was proximal to the peak LOS pressure point in HH subjects but distal to it in controls after the meal (p<0.05). When supine, the postprandial pH transition point crossed the SCJ in those with HH (p=0.03).

Conclusions Healthy volunteers with HH have increased intrasphincteric reflux and lengthening of cardiac mucosa in the absence of traditional transsphincteric reflux.

  • HIATAL HERNIA
  • GASTROESOPHAGEAL REFLUX DISEASE
  • GASTROINTESTINAL MOTILITY

Statistics from Altmetric.com

Significance of this study

What is already known on this subject?

  • Cardiac mucosa at the gastro-oesophageal junction has been shown to lengthen with age and central obesity.

  • Acid damage to distal squamous mucosa can occur by local exposure within the sphincter ‘intrasphincteric reflux’ in the absence of traditional reflux.

  • Hiatus hernia is a key mediator of gastro-oesophageal reflux disease but its significance in the general population is unknown.

What are the new findings?

  • Healthy volunteers with hiatus hernia have lengthening of the cardiac mucosa at the gastro-oesophageal junction.

  • This occurs primarily through intrasphincteric acid exposure rather than by traditional transsphincteric reflux.

How might it impact on clinical practice in the foreseeable future?

  • The expansion of the cardiac mucosa due to intrasphincteric reflux may be relevant to the high incidence of adenocarcinoma at this site in subjects without a history of traditional reflux.

Introduction

Over the past 30 years, there has been an increase in the incidence of adenocarcinoma at the gastric cardia and gastro-oesophageal junction.1–4 In some registries, these adenocarcinomas at the junction between oesophagus and stomach have an incidence similar to those within the oesophagus itself.5 These junctional adenocarcinomas share similar epidemiology to oesophageal adenocarcinoma and in the USA are classified as oesophageal cancers.6 However, unlike oesophageal adenocarcinomas, these cardia and junctional cancers have only a very weak association with a previous history of reflux symptoms.7 ,8 These observations suggest that the mechanism of acid exposure inducing these junctional cancers may be different from traditional transsphincteric reflux.9 ,10

We recently studied the anatomy and physiology of the gastro-oesophageal junction in healthy volunteers without a history of significant reflux symptoms.11 We observed that both increasing age and increasing central obesity were independently associated with proximal extension of the cardiac columnar mucosa. None of these subjects had evidence of increased acid reflux extending across the lower oesophageal sphincter (LOS), however there was evidence of ingress of gastric acid within the sphincter region itself. We concluded that age and central obesity were damaging the integrity of the LOS, allowing gastric acid to penetrate within the sphincter and onto the most distal oesophageal squamous mucosa lying within the sphincteric region. We proposed that this intrasphincteric reflux could explain the proximal extension of the cardiac columnar mucosa arising due to squamocolumnar metaplasia of the most distal oesophagus. As our subjects were healthy volunteers, we also suggested that this process might account for the substantial incidence of adenocarcinoma at the gastro-oesophageal junction in subjects without a history of reflux symptoms.8 ,12

In our previous study, 18% of our healthy volunteers had a hiatus hernia detected by MRI and were excluded from the published analysis. If proximal extension of cardiac columnar mucosa and intrasphincteric acid exposure is indeed due to dysfunction of the LOS caused by age and/or central obesity, then similar changes may occur in subjects with LOS dysfunction due to hiatus hernia. Although there is an extensive literature on the effect of hiatus hernia on LOS function in subjects presenting with reflux symptoms,13–15 there is little on the effects of hiatus hernia detected in healthy volunteers.

In our current study, we have examined the anatomy and physiology of the gastro-oesophageal junction in our healthy volunteers with hiatus hernia and compared them with healthy volunteers without hiatus hernia, matched for age, sex and central obesity.

Methods and materials

As part of our recently published study we recruited 62 healthy volunteers negative for Helicobacter pylori and without known gastro-oesophageal reflux. Fifteen volunteers who completed the full study protocol were found to have a hiatus hernia; 10 diagnosed by MRI, 12 by endoscopy and seven by both modalities. One additional volunteer found to have a hiatus hernia on MRI did not complete the study protocol. The subgroup diagnosed by MRI were excluded from the original obesity study. In this current study, we compared these volunteers who had a hiatus hernia with age-matched and sex-matched controls from the original study cohort.

Study protocol

Study day 1: clinical and MRI assessment

Clinical details were recorded including demographics and anthropometric measures. Volunteers completed a validated survey reporting reflux symptoms.16 MRI scans of the abdomen were performed (Philips 1.5 T MRI scanner, Surrey, UK) before and 45 min after a standardised meal consisting of fried battered fish and chips. A hiatus hernia was identified when the gastro-oesophageal junction was proximal to the diaphragmatic hiatus. Where present on MRI, the length of the hiatus hernia was measured. In addition, the diameter of the gastro-oesophageal junction at the diaphragmatic hiatus was measured in all participants.

Study day 2: endoscopy with biopsies and placement of radio-opaque clip

Volunteers attended after an overnight fast. Upper GI endoscopy was carried out using standard equipment with throat spray or intravenous sedation according to volunteer preference. Anatomy of the upper GI tract was examined and hiatus hernia defined endoscopically as separation of the diaphragmatic indentation and the top of the gastric folds by at least 2 cm. Where hiatus hernia was identified its length was measured and recorded. Biopsies were taken across the squamocolumnar junction (SCJ) in a craniocaudal direction to include both squamous mucosa and glandular mucosa in the same sample. Intraprocedure pathology feedback of biopsy accuracy was available and up to three biopsies were taken to achieve an optimal sample allowing measurement of cardiac mucosal length. Finally, the SCJ was marked by two endoclips (HX-610-135; Olympus, Southend-on-Sea, UK). Biopsies were carefully orientated for histological processing as previously described.11

Study day 3: combined pH and manometry with fluoroscopy

Volunteers attended after an overnight fast. A combined high-resolution manometry and pH probe was passed through the anaesthetised nostril. The pH probe used was customised for purpose and made up of 12 pH sensors (Synectics Medical, Enfield, UK). Manometry recordings were taken using a high-resolution probe comprising 36 solid-state sensors. The probes were combined in a standardised way to allow correlation of pressure and pH findings. Recordings were taken for 15 min with volunteers seated in the upright posture and 15 min supine. Volunteers then consumed the same standardised meal as on study day 1 and were asked to eat until full. After the meal, recordings were continued for 45 min seated upright and a further 30 min supine. Fluoroscopy was performed for 30 s during each phase to allow visualisation of the endoscopically placed clip.

Data analysis

Cardiac mucosal length

The cardiac mucosa was considered ‘fully measurable’ where consecutive squamous, cardiac and glandular mucosal types were present in continuity in the same biopsy. Specimens were considered ‘measurable’ if squamous mucosa and cardiac mucosa were present in continuity within the same biopsy. A third category of ‘not analysable’ was allowed where none of the samples taken had at least two mucosal subtypes in continuity within the same biopsy. Samples defined as ‘not analysable’ were not included in the measurement of cardiac mucosal length.

The cardiac mucosa was defined as columnar epithelium devoid of parietal cells. The proximal demarcation of cardiac mucosa at the SCJ was clear and the distal limit was taken as the appearance of parietal cells. The cardiac mucosal length was measured using an ocular micrometre by two independent GI pathologists (JJG, MHD) blinded to clinical characteristic of the volunteer.

Inflammatory scores

Biopsies were semiquantitatively scored for acute and chronic inflammation and for reactive atypia, a measure of cell turnover and repair, using scores of 0–3 as follows: 0=absent, 1=mild, 2=moderate and 3=severe. Acute inflammation was defined by the presence of polymorphs and chronic by a monocytic infiltrate.

Acid data

To measure intrasphincteric acid position, the mean pH was calculated for each of the 12 sensors for each phase of the study. The first pH sensor recording a mean pH drop of 1 unit was identified. This was termed the pH transition point. We have previously shown that this point represents the abrupt transition from oesophageal to gastric pH.11

The position of this sensor relative to the upper border of LOS, the SCJ and the peak LOS pressure was measured. Distances were corrected for 11 mm spacing of pH sensors.

Acid exposure also was determined at the following points:

  1. The traditional site 5 cm above the upper border of the LOS.

  2. The sensor 11 mm proximal to the sensor detecting the pH transition point.

These sites were defined as detecting traditional and short segment reflux, respectively. Traditional acid reflux was considered present where the pH was below 4 at least 4% of the time at the sensor 5 cm proximal to the upper border LOS and short segment reflux where the pH was <4 at least 4% of the time in the sensor 11 mm proximal to the pH transition point.

Prevalence of the double-peaked pressure profile

For each phase of the study the manometric recording was split into 60 s intervals. For each of these segments the pressure profile of the LOS was examined using the ManoView analysis software and based on both the colour contour plot and the pressure profile. The profile was classified as ‘single’ where one pressure peak was identified and ‘double’ where the LOS comprised two pressure peaks. A third observation of ‘not analysable’ was allowed where there was insufficient stable sphincter tone to enable classification such as occurred with the presence of transient LOS relaxations or frequent swallows. For each group the total time spent in double and single pressure profiles was recorded.

Detailed analysis of LOS

For each of the four periods where fluoroscopic screening was available the LOS was analysed in detail. Data from six consecutive inspirations and six consecutive expirations were selected using a custom-made computer programme. The pressure profile was classified as ‘single or double’ where one or two peaks were seen, respectively.

The position of the upper border LOS was calculated on inspiration as a decrease in pressure to within 2 mm Hg of intragastric pressure moving proximally from the sensors within the LOS to the sensors within the oesophagus. The upper border LOS was used as a local reference point for the positions of the SCJ and pH transition point. Peak LOS pressure was calculated on both inspiration and expiration as the peak pressure within the LOS irrespective of pressure profile. The inspiratory augmentation pressure was calculated by subtracting the expiratory LOS pressure from the inspiratory LOS pressure.

Fluoroscopy

The position of the SCJ was derived from the fluoroscopic images for a median of six inspirations and using the manometre and pH sensors as reference and internal scale. This was calculated relative to the upper border LOS.

Statistical analysis

All results are presented as median and IQR unless otherwise stated. The Mann–Whitney U test was used for comparison between groups and differences were considered significant at a p value of <0.05.

Results

Group characteristics

Of the original study group, 15 volunteers who completed the study (11 male, aged 38–74 years) were found to have a hiatus hernia. Age range in the control group was 28–73 years (p=0.206) and 11 were male. Median body mass index (BMI) in the hiatus hernia group was 25.9 kg/m2 (range 21.0–35.6, IQR 5.8) and waist circumference 91.0 cm (range 77–119, IQR 24). Corresponding values for those without hiatus hernia were 28.7 (16.7–34.5, IQR 7.8) (p=0.650) and 99 cm (70–118 cm, IQR 23) (p=0.820), respectively.

Twelve of the 15 hiatus hernias were diagnosed by endoscopy, 10 by MRI and 7 by both modalities. For those diagnosed by MRI the median length of the hiatus hernia was 2.3 cm (IQR 0.79) and by endoscopy the median length was 3 cm (IQR 0).

MRI measures

The diameter of the diaphragmatic hiatus, measured by MRI was significantly larger in the hiatus hernia group. This was true for both fasting (13.3 mm (IQR 5.1) vs 10.0 mm (IQR 2.75), p=0.001) and for after the meal (14.9 mm (IQR 6.5) vs 9.7 mm (IQR 3.45), p=0.006).

This widened hiatus was seen irrespective of the method by which the hiatus hernia was originally detected. For the 12 volunteers diagnosed with hiatus hernia on the basis of endoscopy the diameter of the hiatus in fasting was 13.6 mm (IQR 5.8) compared with 10.6 mm (IQR 3.2 mm) in those without hiatus hernia on endoscopy (p=0.003). For the 10 volunteers diagnosed by MRI the diameter of the diaphragmatic hiatus was 15.7 mm (IQR 5.2) compared with 10.25 mm (IQR 3.1) for those negative for hiatus hernia by MRI (p<0.001). For the seven volunteers diagnosed those with hiatus hernia confirmed on two modalities the diameter of the hiatus was widened compared with those with hiatus hernia diagnosed on one modality (17 mm (IQR 3.8) vs 12.4 mm (IQR 2.6), p=0.009 for fasting, and 17.4 mm (IQR 4.9) vs 11.1 mm (IQR 5.9), p=0.006 after the meal).

Junctional mucosal characteristics

Three volunteers had Barrett's oesophagus identified endoscopically and were excluded from the analysis of cardiac mucosal length and junctional mucosal characteristics. All three had an associated hiatus hernia. Of the remaining 12 in the hiatus hernia group nine had cardiac mucosa that was either ‘measurable’ or ‘fully measurable’ compared with 13 out of 15 in the control group.

Cardiac mucosa was significantly longer in the group with hiatus hernia compared with controls (3.5 mm (IQR 1.0) vs 2.5 mm (IQR 1.0), p=0.014). In only five of the hiatus hernia group was a full span of the cardiac mucosa obtained in a single biopsy (fully measurable) while 11 out of 15 in the control group had a fully measurable cardia (p=0.038).

Moderate chronic inflammation of the cardiac mucosa did not differ significantly between groups (median score for mononuclear infiltrate in hiatus hernia group 2.0 (IQR 1.0) vs 2.0 (IQR 1.0) in controls, p=0.248). Reactive atypia, a marker of cell turnover and repair, was also similar between the two groups (median score 2.0 (IQR 1.0) in hiatus hernia group vs 2.0 (IQR 1.0) in controls, p=0.748). Acute inflammation was not a prominent feature in either group and did not differ between groups (median score, based on polymorphonuclear infiltrate, 1.0 (IQR 2) for hiatus hernia group and 0 (IQR 1.0) for controls, p=0.116).

Acid reflux

There was no significant excess of reflux symptoms in those with hiatus hernia compared with those without (median reflux scores in hiatus hernia group 7 (IQR 3) vs 5 in controls (IQR 3) p=0.065). Acid exposure measured in the traditional manner (percentage of time pH was below 4 at 5 cm above the upper border LOS) was not a feature in either group and did not differ significantly between groups either before or after the meal (table 1).

Table 1

Acid exposure time (median % time pH<4 (IQR)) in pH sensors located 1.1 cm proximal to the pH transition point and 5 cm proximal to the upper border LOS in healthy volunteers with hiatus hernia and controls

Short segment reflux (percentage of time pH was <4 at 11 mm above pH transition point) was greater in the hiatus hernia group versus controls in the supine posture after a meal (p=0.011, table 1).

LOS pressure and inspiratory augmentation pressure

There was no significant difference in the hiatus hernia group compared with controls in peak LOS pressure either on inspiration or on expiration. However, inspiratory augmentation pressure was diminished in the hiatus hernia group after the meal in the supine posture (table 2).

Table 2

LOS pressure and inspiratory augmentation pressure in hiatus hernia and control groups

Prevalence of double peak on high-resolution manometry

In those with hiatus hernia the double-peaked pressure profile was intermittent rather than fixed being seen on manometry a total of 36.6% of the time across all phases of the study (table 3). Only one volunteer demonstrated a double-peaked profile persistently throughout the recording. In the control group, negative for hiatus hernia by both endoscopy and MRI, the double-peaked pressure profile was also seen intermittently and for 25.4% of total recorded time. Although the double peak was seen in both groups, there was a strong trend to increased prevalence of the double peak in those with hiatus hernia (p=0.053). The prevalence of the double peak for each group across the phases of the study is shown in table 3.

Table 3

Percentage time spent in double-peaked pressure profile for hiatus hernia and control groups

Intrasphincteric characteristics of hiatus hernia versus controls

Relative positions of pH transition point, SCJ and upper border LOS

The pH transition point was determined by the position of the most proximal of the 12 pH sensors showing a mean pH fall of 1 unit. We have previously shown that this sensor represents the abrupt change from oesophageal to gastric pH.11 In those with hiatus hernia, the pH transition point was closer to the upper border LOS in all phases of the study aside from fasting upright (FU) (figure 1) (table 4). In the hiatus hernia group, the SCJ was also more proximally sited within the LOS in FU, postprandial upright and postprandial supine with trend to significance in fasting supine. There were no differences between the groups in the relative positions of the pH transition point and SCJ except after the meal in the supine posture in which the pH transition point was measured 0.31 cm proximal to the SCJ in the hiatus hernia group versus 0.55 cm distal to it in controls (p=0.033). The upper border LOS was located a similar distance from the nares in both groups.

Table 4

Detailed analysis of LOS components in hiatus hernia versus controls showing positions of the upper border LOS, pH transition point, SCJ and first pressure peak

Figure 1

Intrasphincteric characteristics of hiatus hernia versus controls after a meal in the supine posture. LOS, lower oesophageal sphincter; SCJ, squamocolumnar junction.

Relative positions of pH transition point, SCJ and LOS peak pressures

To allow comparison between groups, each having both types of pressure profile, the pH transition point was measured relative to either the proximal pressure peak of a double profile or the isolated peak of a single profile. In the hiatus hernia group, the pH transition point was proximal to this pressure peak after the meal in both upright and supine positions but remained distal throughout in controls (table 4). The differences between the hiatus hernia subjects and controls were most pronounced in the supine posture after the meal and these are illustrated in figure 1.

Intrasphincteric characteristics of hiatus hernia versus controls when both have similar pressure profile

Both the hiatus hernia subjects and controls showed a double-peak pressure profile for a similar proportion of the recording period. We compared the intrasphincteric characteristics of the hiatus hernia versus controls when both showed a single-peak profile during fluoroscopy and again when both showed double-peak profile. As this reduced the recording period available for analysis, we combined the fasting, postprandial, supine and erect data.

Comparison of the single-peaked pressure profile in hiatus hernia and controls

A total of 32 single pressure profiles were detailed during fluoroscopy for the hiatus hernia group and 43 for controls. In the single-peaked pressure profile there was no difference between the groups in the position of the upper border LOS from the nares. However, the SCJ and pH transition point were more proximally sited relative to the nares in the group defined as having a hiatus hernia in the original study protocol (table 5).

Table 5

Analysis of single-peaked pressure profile in hiatus hernia and controls: positions of SCJ, pH transition point and peak LOS pressure relative to nares and to upper border LOS

The pH transition point was closer to the upper border LOS in the hiatus hernia group versus controls. Furthermore, the pH transition point was sited proximal to the peak LOS pressure in those with hiatus hernia but distal to it in controls.

Comparison of the double-peaked pressure profile in hiatus hernia and controls

A total of 28 double-peaked pressure profiles were analysable for the hiatus hernia group versus 16 in controls.

In the double-peaked pressure profile the distance from the nares to the upper border LOS was similar between the groups (table 6). The SCJ and the pH transition point were however proximally displaced in those with hiatus hernia with respect to controls when measured relative to the nares.

Table 6

Analysis of double-peaked pressure profile in hiatus hernia and controls: positions of SCJ, pH transition point and proximal pressure peak relative to nares and upper border LOS

Both the SCJ and the pH transition point were also more proximally located with respect to the upper border LOS. When the position of the pH transition point was measured with respect to the proximal pressure peak of the double peak it was located proximal to it in those with hiatus hernia and distally in controls.

Discussion

The hiatus hernias that were detected in our healthy volunteers were not associated with increased traditional transsphincteric acid reflux or reflux symptoms. However, these subjects with hiatus hernia did have proximal extension of their cardiac mucosa. In addition, the hiatus hernia subjects had short segment acid reflux and their pH transition point located proximal to their SCJ following a meal in the supine position. The proximal extension of the cardiac mucosa is likely to be due to squamocolumnar metaplasia of the distal oesophagus induced by the increased intrasphincteric acid exposure.

Of our 15 healthy volunteers with hiatus hernia, 12 were diagnosed at endoscopy, 10 with MRI scan and seven by both. The variation in findings by these two tests performed at different times is in keeping with the intermittent nature of hiatus hernia.17 ,18 The diameter of the diaphragmatic hiatus assessed by MRI was increased in the hiatus hernia subjects consistent with the disorder being associated with a weakness of the diaphragmatic crura. This widened hiatus detected by MRI was the case irrespective of the method by which the hiatus hernia was originally detected although in those confirmed to have a hiatus hernia by both modalities the widening was most apparent. This may reflect those with a more substantial crural defect spending more time in the hiatus hernia state.

High-resolution manometry has been claimed to be useful in diagnosing hiatus hernia on the basis of demonstration of a double-peak pressure profile due to the separation of the intrinsic and extrinsic sphincter.19 In our subjects with hiatus hernia, the double-peaked pattern was observed intermittently and for <50% of the recording time and this was similar to the subjects who were negative for hiatus hernia by both MRI scan and endoscopy. Scheffer et al 20 reported that a double peak was observed 53.2% of the time in patients with reflux disease compared with only 14.5% of the time in controls. The same group in a further study reported that a double peak was present for a similar period in patients with reflux disease with a small hiatus hernia and controls but was present for a greater proportion of time in patients with reflux disease with a large hiatus hernia.17 The similar prevalence of double peak in our subjects with hiatus hernia versus controls is consistent with them having small hiatus hernias. The explanation for the double peak present in our healthy controls and also those studied by Bredenoord et al is unclear. However, Miller et al 21 described three pressure contributions to the LOS in healthy subjects; one from the crural diaphragm, one attributable to the intrinsic LOS and a third which the authors hypothesised represented the sling and clasp fibres of the proximal stomach, that is, the gastro-oesophageal flap valve.21 The third pressure contribution was absent in the patients with reflux symptoms. It may be therefore that rather than representing an intermittent hiatus hernia, the double peak in our healthy volunteers represents the manometric signature of a healthy gastro-oesophageal flap valve.

There was no evidence of increased acid exposure in our hiatus hernia subjects versus controls when measured at the traditional location 5 cm above the upper border LOS. In addition, our subjects with hiatus hernia did not have increased reflux symptoms. Hiatus hernia is associated with reflux disease13 and the absence of traditional reflux in our healthy volunteers with hiatus hernia might be explained by their well-maintained intrinsic LOS. They had evidence of impaired extrinsic sphincter function as seen in their widened hiatus on MRI and their impaired respiratory augmentation. The lack of reduction of peak LOS pressure in our hiatus hernia subjects despite this clear evidence of impaired extrinsic sphincter function indicates well-maintained intrinsic sphincter function and the latter may be protecting them from transsphincteric reflux. The hiatus hernias in our present study must represent the majority of hiatus hernias in the general population as we detected them in 26% of healthy volunteers by either endoscopy or MRI. The fact that these hiatus hernias were not associated with transsphincteric reflux may explain the controversies which have existed about the relevance of hiatus hernia to reflux disease.22

Though our healthy volunteers with hiatus hernia did not have traditional transsphincteric reflux, they did have evidence of short segment acid reflux detected within the sphincter region at 11 mm proximal to the pH transition point but only following the meal and in the supine position. The occurrence following a meal may be explained by the increased stress on the LOS known to occur following a meal23 and the occurrence only in the supine position due to the fact that hiatus hernias tend to reduce after a meal when the subject is in the erect position.20

Our healthy volunteers with hiatus hernia had abnormalities of their mucosa within the region of the LOS. The length of their columnar cardiac mucosa was greater compared with those without hiatus hernia and this was after full correction for age and central obesity. In addition, the number of volunteers whose mucosa was longer than could be measured by our technique was significantly greater in those with versus without hiatus hernia indicating that the observed difference may be an underestimate. Three volunteers with hiatus hernia had endoscopic evidence of Barrett's oesophagus and were excluded from analysis. In all subjects, the cardiac mucosa was inflamed.

More detailed analysis of the gastro-oesophageal junction was possible during the periods of fluoroscopy when we had accurate information on the relative positions of the SCJ, pH transition point and sphincter manometric landmarks. The distance from the nares to the upper border of the LOS was not significantly different but in the hiatus hernia group the SCJ and pH transition point were closer to the nares and also closer to the upper border of the LOS. After the meal, the pH transition point was also proximal to the first pressure peak in the group with hiatus hernia but distal throughout in controls. These differences were also apparent when the hiatus hernia and controls were compared when showing similar pressure profiles (ie, single or double peak).

How can we explain the observed proximal displacement of the pH step-up point and SCJ relative to the upper border of the LOS and nares without the proximal displacement of the upper border of the LOS relative to the nares? Perhaps, this represents the alteration in the relationship between extrinsic and intrinsic sphincters in those with a small or intermittent hiatus hernia. In the absence of a hiatus hernia where the anatomy at the gastro-oesophageal junction is preserved the extrinsic sphincter overlaps and reinforces the intrinsic sphincter and the pressure inversion point is proximal to the peak sphincter pressure. The upper border of the LOS would be detected as a pressure increase from oesophageal pressure but the pressure contribution could come from the extrinsic sphincter, the intrinsic sphincter or both. Our observations could be consistent with the upper border LOS in normal subjects being formed by the extrinsic sphincter with the intrinsic sphincter positioned distally. Thus, in our subjects with early hiatus hernia there could be proximal movement of the gastro-oesophageal junction (GOJ) including the intrinsic sphincter, SCJ and pH transition point within and even across the extrinsic sphincter. This would only be detectable from the nares once it was persistently proximally displaced with respect to the extrinsic sphincter. Some of the proximal displacement of the SCJ may also be explained by the proximal extension of cardia mucosa and proximal displacement of the pH transition point by intrasphincteric ingress of gastric acid. Finally, the lack of proximal displacement might be partly artefactual due to difficulty in detecting a small change over a relatively long distance.

The lengthening of the cardiac mucosa is likely to be due to the observed acid ingress within the sphincteric region and consequent columnar metaplasia of the distal oesophageal squamous mucosa. Several of the abnormalities observed in our healthy volunteers with hiatus hernia are likely to contribute to the intrasphincteric reflux. The SCJ normally lies within the region of the intrinsic sphincter and closure of the latter prevents gastric juice from refluxing onto oesophageal squamous mucosa. However, maintaining closure of the LOS also depends upon its reinforcement by the extrinsic sphincter and in our hiatus hernia subjects there was weakness of the extrinsic sphincter and intermittent loss of its superimposition relative to the intrinsic sphincter. This impairment of extrinsic sphincter function will predispose to distal opening of the intrinsic sphincter and acid ingress. The observation that the pH transition point in our hiatus hernia subjects was positioned proximal to the peak pressure point within the sphincter is also likely to be important. This will mean that acid secreted by the most proximal gastric mucosa lying above this peak pressure point will tend to flow proximally onto the most distal squamous mucosa rather than distally into the stomach.

These changes observed in subjects with hiatus hernia are similar to those which we observed in subjects without hiatus hernia and associated either with central obesity or increased age. Together, they suggest that stress or ageing or damage to the supporting structures of the gastro-oesophageal junction allow acid to ingress within the sphincteric region encroaching on the most distal, oesophageal mucosa without increased reflux across the sphincter.

This dysfunction of the gastro-oesophageal barrier and associated metaplasia of the distal oesophageal mucosa appears to be common in our population. It occurs in the significant proportion of healthy volunteers who have hiatus hernia and is also observed in healthy volunteers without hiatus hernia but with central obesity or with increasing age. Though the metaplastic change arising at the gastro-oesophageal junction might not carry the high risk of adenocarcinoma associated with transsphincteric reflux and full Barrett's oesophagus, the high prevalence of the condition may contribute to the significant number of cancers occurring at the cardia and junction in subjects without a reflux history.

References

Footnotes

  • Contributors EVR: conduct of study, acquisition of data, analysis and interpretation of data, statistical analysis, preparation and drafting of manuscript. MHD: analysis and interpretation of data, statistical analysis, histological analysis and intellectual contribution. AAW: conduct of study, acquisition of data and intellectual contribution. DRM: intellectual contribution. JJG: histological analysis and intellectual contribution. AWK: data analysis and intellectual contribution. SAB: acquisition and analysis of radiology data and intellectual contribution. KELM: study concept and design, interpretation of data, critical revision of manuscript and study supervision.

  • Funding Western Infirmary Endowment Fund and a research grant (CZB/4/709) from the Scottish Chief Scientist Office.

  • Competing interests None declared.

  • Patient consent Obtained.

  • Ethics approval West of Scotland Ethics Committee.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.