- Top of page
- Materials and methods
Background : There are conflicting reports on the role of gastro-oesophageal reflux disease (GERD) and Helicobacter pylori infection in the aetiology of carditis.
Aim : The role of reflux and H. pylori infection in causing carditis was assessed in 113 consecutive patients with GERD and in 25 controls.
Methods : All subjects underwent endoscopy and pH test and carditis was diagnosed on biopsies taken across the squamocolumnar junction. Helicobacter pylori was assessed by histology and rapid urease test. GERD was diagnosed by endoscopic oesophagitis or abnormal pH test.
Results : Carditis was detected in 53 of 71 GERD patients and in 15 of 20 controls. Among patients, 18 showed absent, 39 mild and 14 marked cardia inflammation and their H. pylori infection rates were 17, 23 and 57%, respectively (P < 0.025). Most patients with carditis (68%) lacked H. pylori infection. pH-metry was abnormal in 15 of 18 patients with normal cardia, 33 of 39 with mild carditis and 12 of 14 with marked inflammation.
Conclusions : Carditis is a frequent finding in GERD and controls. Mild, non-active carditis is frequent in GERD patients. Marked inflammation is associated with both H. pylori and abnormal pH testing. Thus, both GERD and H. pylori infection may play a role in inducing carditis.
- Top of page
- Materials and methods
In recent years, the incidence of adenocarcinoma at the gastro-oesophageal junction has been increasing.1 These junctional cancers appear to arise from foci of intestinal metaplasia that develop either in the distal oesophagus or in the proximal stomach (the gastric cardia).2
This lesion is usually a consequence of chronic inflammation, and it is logical to assume that intestinal metaplasia at the gastro-oesophageal junction develops as a result of chronic inflammation in the epithelia that normally lines the junction region.3
In this area, three epithelial types are normally found: stratified squamous epithelium, cardia epithelium and gastric oxyntic epithelium. In the oesophagus, gastro-oesophageal reflux disease (GERD) may cause inflammation and reflux oesophagitis can lead to intestinal metaplasia (Barrett's oesophagus), that in its turn predisposes to oesophageal adenocarcinoma.4 In the gastric antrum and body Helicobacter pylori infection causes chronic inflammation and this condition is associated with the development of intestinal metaplasia and cancer.5 It is not clear how gastro-oesophageal reflux, H. pylori infection and other factors could contribute to inflammation and metaplasia in the gastric cardia.
Several recent studies have given contradictory results on the pathogenetic factors favouring inflammation of cardia epithelium.6–19 The conflicting findings published in medical literature on this topic are due to various aspects: differences in patient selection, definition of cardia epithelium, sampling of gastric cardia, lack of a control group, prevalence of H. pylori infection among various countries. One of the main reasons for the contrasting results lies in the fact that the diagnosis of GERD is frequently based only on symptoms or endoscopic lesions, while nowadays we know that the majority of reflux patients do not show endoscopically visible oesophagitis.20 In these cases, 24-h oesophageal pH monitoring remains the most sensitive diagnostic tool, although it cannot be considered infallible.21 In the few studies in which histological results were compared with pH measurements, carditis was found to be associated with an increased oesophageal acid exposure.15, 19
Therefore, we performed this prospective study in order to evaluate the respective roles of oesophageal acid exposure and H. pylori infection in the pathogenesis of inflammation at the squamocolumnar junction (SCJ) or carditis. We compared patients with typical and atypical symptoms of gastro-oesophageal reflux and subjects belonging to an appropriate control group; in both groups upper gastrointestinal (GI) endoscopy, histology and 24-h oesophageal pH monitoring were performed.
Materials and methods
- Top of page
- Materials and methods
This study was conducted in accordance with the Declaration of Helsinki and its revisions and was approved by our local Ethics Committee. All patients were asked to give written informed consent before participating into the study. Patients with severe systemic diseases, active peptic ulcer, previous GI surgery, previous attempts of H. pylori eradication, use of aspirin or non-steroidal anti-inflammatory drugs, women who were pregnant or with breastfeeding, alcohol or drug abuse, oesophageal carcinoma and severe psychiatric disease were excluded from the study.
A total of 113 consecutive patients (62 males and 51 females; mean age 51.9 ± 14.4 years) with typical and atypical symptoms of gastro-oesophageal reflux, who agreed to undergo both upper endoscopy and oesophageal pH monitoring, were recruited for this investigation. They were referred to our Gastroenterological Unit from July 1999 to July 2003. The diagnosis of GERD was based on the history of typical symptoms, the endoscopic evidence of oesophagitis and/or abnormal oesophageal pH testing, which was mandatory in patients with atypical symptoms. Typical symptoms consisted in the presence of heartburn and/or acid regurgitation at least twice a week for the previous 4 weeks; atypical symptoms were instead recurrent chest pain of more than 3 months duration with cardiologist's evaluation that symptoms were not cardiac in origin or recurrent respiratory symptoms for at least 3 weeks (chronic cough, asthma and hoarseness) with absence of concurrent infectious or allergic causes. All patients underwent upper GI endoscopy after an overnight fast. The proximal edge of the gastric folds was identified as the SCJ or Z-line between the oesophagus and the stomach and in the case of irregular, tongue-like Z-line the proximal limit of the extension of the tongues was considered as the proximal limit of the junction. In patients with hiatal hernia the SCJ was localized at the proximal edge of the hernia sac, corresponding to the proximal margin of the gastric folds. The endoscopic examinations were always performed by two operators (CM and PD) and biopsies were obtained with the endoscope in the conventional antegrade position. Three specimens were taken from the antrum and the corpus of the stomach and two across the SCJ, and at 2 and 4 cm above it. Cardia-type epithelium was defined on the basis of the coexistence of both squamous and glandular epithelium in the same sample and showed either purely antral-like mucus glands or an admixture of mucus and oxyntic glands. The presence and degree of inflammation, activity (i.e. intra or periepithelial neutrophils) as well as intestinal metaplasia were assessed on haematoxylin-eosin and Alcian Blue/periodic acid Schiff-stained slides. The diagnosis of H. pylori infection was based on the concomitant positivity of histology (modified Giemsa stain) and CLO-test (Delta West Ltd., Bentley, Australia), which was performed on both body and antrum biopsies. We used the Los Angeles classification for grading oesophagitis22 and the updated Sydney classification for grading gastritis.23 Histological assessment was performed by two expert pathologists (RF and LM), independently and in a blinded manner. The degree of inflammation and activity in the cardia was scored from 0 (absent) to 2 (marked) and the agreement between the two above pathologists was high (k = 0.78). The diagnosis of Barrett's oesophagus was based on the histological detection of specialized intestinal metaplasia on biopsy samples taken from the distal oesophagus, when an endoscopically apparent columnar epithelium of 3 cm or less was found.
Furthermore, every patient underwent 24-h oesophageal pH-metry. Oesophageal pH was measured by a glass minielectrode (Ingold 440 M 3, Urdorf, Switzerland) positioned 5 cm above the upper border of the lower oesophageal sphincter, which was manometrically identified prior to each pH study. The electrode was connected to a datalogger that permitted pH readings to be taken every 6 s (Proxima Light 2; Synectics AB, Stockolm, Sweden). At the beginning and at the end of each examination, the pH glass electrode was calibrated at 37 °C, using buffer solutions of pH 7.0 and pH 1.0 (Merck, Darmstadt, Germany). During the test day, meal time and composition were standardized.24 Antisecretory and prokinetic drugs, if used by the patients, were discontinued at least 4 weeks before the examination. In case of reflux symptoms during the wash-out period patients were asked to take only antacid medications.
The reflux parameters were assessed according to Johnson-DeMeester.25 Of them, only the percentage total time spent at pH < 4.0 units during the whole 24-h period was evaluated in our study, because it is the most useful discriminator between physiological and pathological reflux.21 In our experience, pH testing was considered abnormal if pH < 4.0 units was present for more than 5.5% of the total 24-h time.26 Oesophageal pH monitoring was always performed by the same physician (PZ).
The control group consisted of 25 individuals (10 males and 15 females; mean age 58.4 ± 13.1 years) referred to our open-access gastroenterological service for symptoms of non-oesophageal origin, who did not show any endoscopic lesion in the upper digestive tract and had normal 24-h oesophageal acid exposure evaluated by pH test.
The statistical analysis was carried out by means of chi-square and Fisher's exact tests, when appropriate; the level of significance was set at P < 0.05.
- Top of page
- Materials and methods
The demographic characteristics of patients with carditis and controls are reported in Table 1. Cardia epithelium was identified in 71 of 113 patients (63%) and in 20 of 25 controls (80%) while in the other cases Z-line biopsies yielded only squamous epithelium. Cardia inflammation was found in 53 of the above 71 patients (75%) and in 15 of 20 controls (75%).
Table 1. Demographic characteristics of reflux patients and controls with or without carditis
| ||Age (median and range)||M/F||Alcohol (%)||Smokers (%)||Coffee consumers (%)||Hiatus hernia (%)|
| Normal cardia (n = 18)||54 years (22–73)|| 6/12||7/18 (39)||5/18 (28)||9/18 (50)||7/18 (39)|
| Carditis (n = 53)||54 years (22–73)||29/24||22/53 (42)||18/53 (34)||25/53 (47)||27/53 (51)|
| Normal cardia (n = 5)||47 years (31–70)|| 0/5||2/5 (40)||2/5 (40)||3/5 (60)||2/5 (40)|
| Carditis (n = 15)||63 years (20–84)|| 7/8||6/15 (40)||4/15 (27)||10/15 (67)||6/15 (40)|
Among cases with cardia epithelium, H. pylori infection was observed in 20 of 71 patients (28%) and in eight of 20 controls (40%), with no significant difference between the two groups (P = 0.41). Carditis was more frequent in cases with (25/28; 89%) than in those without (43/63; 68%) H. pylori infection (P = 0.038) (Table 2). Among the 71 reflux patients with cardia epithelium, 18 showed absent, 39 mild and 14 marked inflammation. The respective H. pylori infection rates were three of 18 (17%), nine of 39 (23%) and eight of 14 (57%) (P < 0.025; Table 3). Despite the above correlation between the severity of inflammation and the occurrence of H. pylori infection, it is noteworthy that most patients with carditis (68%) lacked infection. Both antrum and body gastritis were present in all the patients in whom carditis was associated with H. pylori infection.
Table 2. Prevalence of Helicobacter pylori infection in gastro-oesophageal reflux disease (GERD) patients with and without carditis and in controls
| ||H. pylori+ve (%)||H. pylori−ve (%)|
| Carditis (n = 53)||17 (32)||36 (68)|
| Normal cardia (n = 18)||3 (17)||15 (83)|
| Carditis (n = 15)||8 (53)||7 (47)|
| Normal cardia (n = 5)||0 (0)||5 (100)|
Table 3. Prevalence of Helicobacter pylori infection, abnormal oesophageal acid exposure and degree of carditis in patients with cardia epithelium
|Carditis||H. pylori−ve (%)/ pH-metry+ve (%)||H. pylori+ve (%)/ pH-metry+ve (%)|
|Absent (n = 18)||15 (83)/13 (72)||3 (17)/2 (11)|
|Mild (n = 39)||30 (77)/24 (62)||9 (23)/9 (23)|
|Marked (n = 14)|| 6 (43)/6 (43)||8 (57)/6 (43)|
Abnormal 24-h pH-metry was found in 95 of 113 GERD patients (84%) and in 60 of the 71 patients (85%) with cardia epithelium. As shown in Table 3, the prevalence of abnormal oesophageal acid exposure was not significantly different among patients with normal cardia (15/18; 83%), those with mild (33/39; 85%) and marked carditis (12/14; 86%). An altered oesophageal acid exposure was slightly more frequently found in H. pylori-negative patients (68/78; 87%) than in H. pylori-positive patients (27/35; 77%) (P = 0.26 ns), whereas there was no difference in pH testing between H. pylori-negative and H. pylori-positive patients with marked cardia inflammation. A normal pH test was found in eight of 53 (15%) patients with carditis and in three of 18 (17%) patients with normal histological findings of cardia epithelium. Table 4 confirms that an abnormal pH test was found in most patients with and without carditis, who were both H. pylori-positive and negative.
Table 4. Percentages of normal and abnormal pH tests in relation to Helicobacter pylori status in patients with and without carditis
| ||pH-metry+ve (%)||pH-metry−ve (%)|
| H. pylori+ve (n = 17)||88||12|
| H. pylori−ve (n = 36)||83||17|
| H. pylori+ve (n = 3)||67||33|
| H. pylori−ve (n = 15)||87||13|
Cardia inflammation was detected in 45 of 60 (75%) GERD patients with an abnormal oesophageal acid exposure and only 15 of 45 (33%) were H. pylori-colonized. However, inflammation of cardia epithelium was associated with H. pylori infection in 17 of 53 (32%) patients and 45 of the same patients (85%) had contemporarily an abnormal oesophageal pH test. Finally, seven of 12 H. pylori-negative controls showed carditis (Table 2) that cannot be explained either on the basis of reflux or H. pylori infection.
Active inflammation (i.e. neutrophils) was uncommon in cardia mucosa of GERD patients (nine of 71, 13%); it was associated with H. pylori infection in six cases, and with abnormal oesophageal pH test in seven cases.
In our series 11 of 71 patients (15%) showed intestinal metaplasia in the cardia epithelium and seven were men. The phenotype of intestinal metaplasia was incomplete in nine patients, complete in one and mixed in one. It is noteworthy that all these patients had an abnormal oesophageal acid exposure, except the patient with complete intestinal metaplasia, while only four of 11 were infected with H. pylori. Three patients had non-erosive reflux disease with histological oesophagitis and eight patients had macroscopic oesophagitis, with Barrett's epithelium in one of them.
- Top of page
- Materials and methods
We investigated 113 patients with typical and atypical symptoms of gastro-oesophageal reflux and Z line-targeted biopsies yielded cardia epithelium in 71 of them (63%) while only squamous epithelium was obtained in other 42 cases. In the control group, the occurrence of cardia epithelium was similar (80%). This finding confirms the technical difficulty in obtaining cardia epithelium in biopsy specimens, as already noticed by others.15, 27 Cardia is a short epithelial segment whose length varies from 1 to 4 mm 28, 29 below the SCJ, that is much less than the 1 or 2 cm usually considered as normal;30 this explains why it can be easily missed even in experienced hands. Lembo et al.31 have also found that cardia lesions can be greatly underestimated when sampling is limited to the mucosa 1–2 cm below the SCJ.
Some authors showed that cardia epithelium is not found in a number of patients and, more importantly, when it is found, is frequently inflamed14, 15, 19, 32. This led some of them14, 15, 19 to the conclusion that cardia epithelium is not a normal feature present from birth, but it is acquired as a consequence of abnormal reflux in the distal oesophagus. This hypothesis is controversial and several studies33–35 have clearly shown that cardia epithelium represents a normal structure of gastric mucosa as it has been found in autopsy studies performed on the gastro-oesophageal junction of embryos, fetuses, infants, young children and adults.34, 35
Our study shows that carditis is very frequent in GERD patients in whom cardia epithelium has been successfully biopsied (75%), but it is also found in controls with the same frequency (75%). This suggests that cardia inflammation is very common and other studies have shown that there is no significant difference in the prevalence of carditis between dyspeptic or GERD patients and controls9, 11. It must be emphasized that our study was the only one showing that carditis in the control group is associated with a normal oesophageal pH test and this result challenges the conclusions of some authors19 that cardia mucosa represents a marker of gastro-oesophageal reflux. This high prevalence of carditis in different populations may be due to the fact that carditis may be consequent to causes other than reflux and H. pylori infection, such as physiological reflux (wear and tear) or the mechanical trauma in the transitional zone.36 Katzka et al.37 have recently shown in 10 normal volunteers that cardia acid exposure is present at a level intermediate between that of the oesophagus and stomach, but it may persist for longer periods of time. Recently, other authors38 have also proposed bile reflux as a cause of carditis.
In our series of GERD patients carditis was a frequent finding (75%) and most cases (68%) were not associated with H. pylori infection. Cardia inflammation occurred in 75% of GERD patients with an abnormal oesophageal acid exposure and only 33% of them were infected by H. pylori. Among 53 patients with cardia inflammation, H. pylori infection was found in 32%, while in the same patients the prevalence of an abnormal oesophageal pH test was 85%. On the contrary, no significant correlation was found between the severity of carditis and the frequency of abnormal oesophageal pH test. The possible causal relationship between carditis and GERD is confirmed by the fact that carditis is frequently (68%) associated with a normal distal stomach.
Studies on the aetiology of carditis have yielded contradictory results. Goldblum et al.9 examined biopsy specimens from the gastric cardia of 58 patients with GERD and 27 control subjects. Carditis was not associated with GERD, because it was found with similar prevalence in patients and controls, while H. pylori infection was present in 22 of 23 patients with GERD and in 11 controls with carditis. The investigators concluded that carditis was associated more with H. pylori infection than with GERD. Other authors8 found cardia epithelium in 116 of 155 unselected patients and carditis was found in 92% of them. The patients with and without this inflammation did not differ significantly in the frequency of symptoms and signs of GERD, but the authors found a strong correlation between carditis and H. pylori infection. Similar findings (i.e. association with H. pylori infection and no correlation with reflux symptoms and endoscopic oesophagitis) were reported by Villani et al.,39 who found also a significant correlation between carditis and retroxyphoid pyrosis in a study involving 190 patients with dyspeptic and/or reflux symptoms. Interestingly these authors reported the same prevalence of carditis as in the present study (75%). At variance with these studies, Oberg et al.15 observed cardia epithelium in 74% of the biopsies they performed at the gastro-oesophageal junction; carditis was found in 96% of these patients and correlated with abnormal reflux. Other authors40 studied 86 unselected patients and concluded that intestinal metaplasia of the cardia is associated with GERD and not with H. pylori infection. Finally, Csendes et al. observed that the presence of carditis and intestinal metaplasia of the cardia was associated with GERD in two different studies.4, 41
It seems clear that patient selection is fundamental to explain the controversies about the aetiology of carditis. Some authors9, 11, 31, 41 have analysed only GERD patients, others36 have evaluated only H. pylori colonized patients and others6, 10, 13, 18, 32, 39, 42 a mixed population undergoing endoscopic examination. Moreover, in many studies11–13, 18, 32, 39, 42, 43 GERD was diagnosed only on the basis of symptoms or endoscopically visible erosive oesophagitis and, nowadays, it is well known that the majority of GERD patients do not have macroscopic oesophageal lesions. This greatly reduces the diagnostic sensitivity of endoscopy in GERD and in our study, in fact, only 42% of GERD patients had macroscopic oesophagitis; a similar result was found by other authors20 who studied a greater group of patients. We used a more sensitive marker of reflux, such as 24-h oesophageal pH monitoring; this corroborates our findings and, more importantly, it contributes to assess whether patients with carditis associated with H. pylori infection also have an abnormal acid exposure. This feature, for instance, was observed in almost all our patients with severe chronic carditis and we would have attributed these cases only to H. pylori infection, if pH testing had not been performed. It is likely that the concomitance of reflux and H. pylori infection occurred in other studies, in which carditis was exclusively related to the germ without the support of an oesophageal pH test.
There is a group of studies in which experimental findings were such that carditis could be clearly related to two different aetiologies. Voutilainen et al.42 showed that carditis was associated with H. pylori infection or with GERD in two different patterns; the former was associated with older age and chronic gastritis and caused by H. pylori infection, while the latter was combined with younger age and the presence of oesophagitis. In the study by Bowrey et al.44 carditis was associated with H. pylori in one-third of patients and was not associated with the infection in the remaining two-thirds. This reflects the prevalence of H. pylori infection in GERD patients, which is of about 30%45 and this low rate is another factor which strongly reduces the contribution of this germ to cardia injury.31 The authors conclude that carditis is likely to recognize two aetiological factors, but it cannot be excluded that an abnormal oesophageal acid exposure would have been found also in the H. pylori-related small group of patients, if a pH test was carried out. In a recent study43 Goldblum et al. concluded that carditis is likely due to a number of different mechanisms, although they found that inflammation of gastric cardia was strongly correlated with H. pylori infection.
Helicobacter pylori infection showed a low prevalence in our GERD patients (28%). Our findings confirm that patients with the most severe inflammation of cardia mucosa have the highest prevalence of H. pylori infection.31 It is noteworthy that almost all our patients with severe carditis had both H. pylori infection and abnormal pH test, and the degree of inflammation does not allow us to predict one of the two above aetiologies. On the contrary, Der et al.46 found that severe inflammation in the cardia was strongly associated with acid reflux, while neutrophilic activity was related to distal gastritis and H. pylori. So, the finding of severe inflammation and neutrophils in carditis may be predictive of the co-existence of GERD and H. pylori gastritis.
In our patients with H. pylori infection and chronic active carditis, the inflammation of both antrum and corpus mucosa was a constant feature. This finding has been already observed in other studies44 and confirms that cardia inflammation is very frequent in patients with antritis due to H. pylori infection.6 In contrast, patients with carditis related to reflux generally show no inflammation of gastric body and antrum.31, 44
We found cardia intestinal metaplasia in 11 of our 71 GERD patients with cardia epithelium (15%). This prevalence is similar to that found in other studies29, 47 and our results seem to suggest that this lesion could be related to GERD. In fact, while H. pylori infection seems not to correlate with the presence of cardia intestinal metaplasia, all but one patients presenting this histological change also showed an abnormal oesophageal pH test. However, no association has been shown between adenocarcinoma of cardia and H. pylori infection, while most adenocarcinomas developing in this area appear to be related to Barrett's epithelium.6
The only way to obtain reliable objective results about the relationship between carditis and GERD or H. pylori infection or even other possible causes is to assess the reversibility of inflammation, once eliminating the above aetiological factors.48 Although it has been shown that H. pylori eradication has dramatically improved inflammation in the gastric cardia in a small group of H. pylori-positive patients,49 no data are available in GERD patients undergoing acid suppressive therapy, and this is a priority which needs to be addressed.
In conclusion, our study shows that carditis is a frequent finding in both GERD patients and in controls. Although both the presence and severity of carditis correlate significantly with H. pylori infection, a high number of carditis occurs in the absence of H. pylori infection in both GERD patients and controls. Mild, non-active cardia inflammation is usually found in patients with abnormal oesophageal pH testing, while patients with severe carditis show both H. pylori infection and altered oesophageal acid exposure. However, some carditis lack both GERD and H. pylori association. On the basis of the above data, it can be asserted that the pathogenesis of inflammation at the SCJ is likely due to multiple factors; accordingly, both abnormal gastro-oesophageal acid exposure and H. pylori infection may play a role in inducing carditis.