Gastrointestinal manifestation of systemic sclerosis—thickening of the upper gastrointestinal wall detected by endoscopic ultrasound is a valid sign

Abstract

Objective. Gastrointestinal (GI) tract involvement has been observed in the majority of patients with SSc. This has been attributed to an accumulation of extracellular matrix within the GI walls. We visualized the walls of the oesophagus, stomach and duodenum with its layers and measured the thickness in SSc patients and control patients utilizing endoscopic ultrasound (EUS).

Methods. Twenty-five SSc patients and 25 controls were evaluated. In addition to analysis of clinical symptoms, endoscopy and EUS (20-MHz miniprobe) were performed. The thickness of the complete wall was measured, and the mucosa, submucosa and muscularis were evaluated separately.

Results. Clinical symptoms of SSc patients were dysphagia (14/25) and heartburn (19/25). Endoscopic findings were hiatal hernia (16/25), oesophagitis (6/25), amotility (19/25) and a dehiscent pylorus (15/25). In comparison with controls, SSc patients had significantly thicker oesophageal [SSc 1.619 (0.454) mm, control 1.392 (0.333) mm; P = 0.025], antral [SSc 1.876 (0.635) mm, control 1.599 (0.291) mm; P = 0.029] and duodenal [SSc 1.730 (0.522) mm, control 1.525 (0.222) mm; P = 0.039] walls. Predominantly, submucosa and muscularis were significantly thicker in SSc patients. The presence of dysphagia or amotility was significantly associated with the thickening of the GI walls.

Conclusions. The EUS revealed a significant thickening of the walls of the upper GI tract in SSc patients. Predominantly, the submucosa and muscularis are enlarged. These results strengthen the hypothesis that increased matrix deposition is an important aspect in the pathogenesis of GI involvement in SSc.

Endosonography, Systemic sclerosis, Gastrointestinal disorder

Introduction

SSc is characterized by vascular lesions due to endothelial cell damage and variable degrees of extracellular matrix accumulation causing fibrotic degenerative changes in organs by proliferation of subendothelial connective tissue [1, 2].

An incidence of gastrointestinal (GI) tract involvement gradable by the Medsger Disease Severity Scale [3] has been observed in ∼80% of the SSc patients of the German nationwide cohort, with the oesophagus as most often the affected segment [4].

Endoscopic ultrasound (EUS) allows an accurate visualization of the GI wall [5, 6]. We therefore proposed that accumulation of the extracellular matrix could be detected by the EUS in SSc patients suffering from GI involvement. To test this hypothesis, the thickness of walls and layers of the oesophagus, stomach and duodenum were measured by the EUS in SSc patients, as well as in non-SSc patients. In addition, we wanted to explore whether differences were detectable between the subgroups of SSc patients.

Methods

Twenty-five SSc and 25 non-SSc patients were included in the analysis. The diagnosis and subclassification of SSc had been confirmed using the clinical presentation according to the ACR criteria [1, 4]. Considering the mixed nature of cardiorespiratory involvement, clinical information about coronary heart disease was recorded [7]. Patients with discomfort of the upper GI tract requiring upper endoscopy were used as the control group. These patients had to fulfill the following criteria: (i) oesophagus: absence of dysphagia and heartburn, macroscopically healthy-looking mucosa at the location of measurement, absence of portal hypertension; (ii) stomach: absence of abdominal weight loss (>5 kg in the past 4 weeks), ulcers at the location of measurement, portal hypertension, macroscopically visible gastritis; and (iii) duodenum: absence of abdominal weight loss (>5 kg in the past 4 weeks), ulcers and portal hypertension, macroscopically healthy-looking mucosa. The study was approved by the Clinical Research Ethics Committee of the University of Regensburg, Germany. All patients gave informed consent to the study according to the Declaration of Helsinki. No patients had to be withdrawn from the study.

Endoscopy was performed using a GIF Q 160 endoscope (Olympus, Hamburg, Germany). The thickness of the submucosa, muscularis and wall of oesophagus, corpus, antrum and duodenum were determined three times in each patient using a 20-MHz miniprobe (Fujinon, Willich, Germany) (supplementary Figure 1, available as supplementary data at Rheumatology Online) and the EUB 8500 (Hitachi, Wiesbaden, Germany). The mean value was calculated and the thickness of the mucosa was determined by subtraction (mucosal thickness = wall thickness − submucosal thickness − muscularis thickness). In both groups, ultrasound measurements were performed at representative areas without visible abnormality. For statistical analysis, the Mann–Whitney U-test was applied (SPSS, Chicago, IL, USA). A P-value of <0.05 was considered as statistically significant.

Results

Study cohort

Twenty-five patients with SSc and 25 controls were enrolled. The female-to-male ratio was 21:4 in SSc and 18:14 in control patients. The mean age was 55.5 (12.6) years in SSc and 57.3 (13.7) years in control patients. Characteristics of SSc patients are listed in Table 1.

Table 1

Clinical data of the SSc population

	Total	dcSSc	lcSSc
Number of patients	25	14	11
Female	21	2	19
Male	4	4	0
ANA positive	24	13	11
Anti-Scl70 positive	11	8	3
ACA positive	11	1	10
RP	23	12	11
Hiatal hernia	16	7	9
Gastritis	15	9	6
Oesophagitis	6	5	1
Dehiscent pylorus	15	7	8
Duodenitis	3	1	2
Dysphagia	14	10	4
Heartburn	19	12	7
Stomach discomfort	7	4	3
Amotility GI tract	19	10	9
Digital ulcers	9	2	7
Weakness of muscles	3	3	0
Atrophia	4	4	0
Proteinuria	1	1	0
Renal crisis	1	1	0
Reduced vital capacity	7	7	0
Heart involvement	10	5	5
Coronary heart disease	1	1	0
Continuous proton pump inhibitor treatment	14	8	6

Clinical symptoms, endoscopic and EUS findings of the oesophagus

Clinical symptoms were dysphagia (14/25) and heartburn (19/25). Endoscopic findings were hiatal hernia (16/25), oesophagitis (6/25) and amotility (19/25). EUS revealed a significantly thicker oesophageal wall [1.619 (0.454) mm] in SSc patients compared with controls [1.392 (0.333) mm] (P = 0.025). The oesophageal wall in SSc patients with dysphagia [1.697 (0.381) mm] was thicker than in SSc patients without dysphagia [1.514 (0.47) mm]. The thickness of the oesophageal wall of SSc patients without dysphagia and controls did not differ significantly. SSc patients with dysphagia also had a significantly thicker oesophageal muscularis [0.609 (0.125) vs 0.530 (0.107) mm; P = 0.021] and oesophageal submucosa [0.447 (0.253) vs 0.348 (0.099) mm; P = 0.044] than controls. SSc patients with GI Medsger Disease Severity 1 had a significantly thicker oesophageal submucosa compared with SSc patients with GI Medsger Disease Severity 0 [0.403 (0.231) vs 0.393 (0.070); P < 0.001] [3]. No significant differences of the wall and its layers were observed between dcSSc and lcSSc patients (Table 2), and between patients with short vs long disease duration (data not shown). Patients with RP not requiring vasodilators tend to have a thinner oesophageal wall and layers compared with patients with digital tip ulcers, though this did not reach statistical significance (data not shown).

Table 2

Clinical symptoms, endoscopic and EUS findings of the oesophagus, the stomach and the duodenum in SSc patients with and without dysphagia, in dcSSc, lcSSc patients and in controls

	SSc with dysphagia	SSc	dcSSc	lcSSc	Controls
Clinical symptoms
Dysphagia	14/14	14/25	10/14	4/11	0/25
Heartburn	12/14	19/25	12/14	7/11	0/25
Stomach discomfort	5/14	7/25	4/14	3/11	6/25
Endoscopy
Hiatal hernia	11/14	16/25	7/14	6/11	3/25
Oesophagitis	3/14	6/25	5/14	1/11	0/25
Amotility	11/14	19/25	10/14	9/11	0/25
Gastritis	8/14	15/25	9/14	6/11	14/25
Dehiscent pylorus	5/14	15/25	7/14	8/11	1/25
Duodenitis	1/25	3/25	1/14	2/11	0/25
Endosonography, mean (s.d.), mm
Oesophagus mucosa	0.597 (0.167)	0.583 (0.273)	0.672 (0.295)	0.531 (0.225)	0.554 (0.226)
Oesophagus submucosa	0.447 (0.253)	0.401 (0.201)	0.425 (0.261)	0.373 (0.098)	0.348 (0.099)
Oesophagus muscularis	0.609 (0.125)	0.601 (0.205)	0.578 (0.141)	0.628 (0.267)	0.530 (0.107)
Oesophagus complete GI wall	1.697 (0.381)	1.619 (0.454)	1.675 (0.440)	1.540 (0.416)	1.392 (0.333)
Corpus mucosa	0.698 (0.276)	0.653 (0.233)	0.633 (0.220)	0.656 (0.255)	0.576 (0.274)
Corpus submucosa	0.461 (0.126)	0.459 (0.137)	0.445 (0.144)	0.477 (0.131)	0.455 (0.214)
Corpus muscularis	0.871 (0.430)	0.780 (0.349)	0.820 (0.430)	0.729 (0.213)	0.734 (0.206)
Corpus wall	2.011 (0.639)	1.881 (0.543)	1.883 (0.611)	1.879 (0.474)	1.844 (0.545)
Antrum mucosa	0.724 (0.408)	0.654 (0.136)	0.612 (0.214)	0.685 (0.415)	0.608 (0.196)
Antrum submucosa	0.436 (0.120)	0.444 (0.128)	0.428 (0.090)	0.464 (0.166)	0.344 (0.077)
Antrum muscularis	0.815 (0.445)	0.772 (0.354)	0.781 (0.426)	0.760 (0.256)	0.647 (0.157)
Antrum wall	2.001 (0.804)	1.876 (0.635)	1.856 (0.632)	1.903 (0.670)	1.599 (0.291)
Duodenum mucosa	0.761 (0.304)	0.643 (0.303)	0.688 (0.310)	0.580 (0.297)	0.518 (0.171)
Duodenum submucosa	0.444 (0.135)	0.427 (0.125)	0.411 (0.135)	0.450 (0.113)	0.391 (0.117)
Duodenum muscularis	0.683 (0.253)	0.676 (0.252)	0.675 (0.271)	0.610 (0.236)	0.616 (0.143)
Duodenum wall	1.888 (0.578)	1.730 (0.522)	1.506 (0.618)	1.668 (0.367)	1.525 (0.222)

Clinical symptoms, endoscopic and EUS findings of the stomach

Stomach discomfort was reported by 7/25 SSc patients and by 6/25 controls. Endoscopic features were the presence of gastritis (15/25 SSc patients, 14/25 controls) and a dehiscent pylorus (15/25 SSc patients and 1/25 controls). There were no significant differences in the thickness of the wall or the layers of the gastral corpus between controls and SSc patients with or without dysphagia. But patients with GI Medsger Disease Severity 1 had a significantly thicker corpus wall and muscularis compared with patients with GI Medsger Disease Severity 0 [1.907 (0.634) vs 1.808 (0.069) mm; P < 0.001 and 0.723 (0.087) vs 0.724 (0.404); P < 0.001, respectively].

The antral wall in SSc patients [1.876 (0.635) mm] and in the subgroup with dysphagia [2.001 (0.804) mm] was significantly thicker compared with controls [1.599 (0.291) mm] (SSc vs controls, P = 0.029; SSc with dysphagia vs controls, P = 0.014). The antral muscularis [SSc 0.772 (0.354) mm, controls 0.647 (0.157) mm; P = 0.06] and the antral submucosa [SSc 0.444 (0.128) mm, controls 0.344 (0.077) mm; P = 0.003] were significantly thicker in SSc patients than in controls. SSc patients with dysphagia had the thickest gastral layers. SSc patients with GI Medsger Disease Severity 1 had a significantly thicker antral wall compared with SSc patients with Medsger Disease Severity 0 [1.966 (0.634) vs 1.593 (0.236) mm; P < 0.001].

Again, there were no significant differences of the gastric wall and its layers between dcSSc and lcSSc patients (Table 2), and between patients with short vs long disease duration (data not shown). In addition, no significant differences were observed between the corpus and antral wall thickness values in each individual patient (data not shown).

Patients with RP not requiring vasodilators tend to have a thinner gastral wall and layers compared with patients with digital tip ulcers, though this did not reach statistical significance (data not shown).

Clinical symptoms, endoscopic and EUS findings of the duodenum

No specific clinical symptoms were identified. SSc patients (3/25) presented with duodenitis. The duodenal wall was significantly thicker in SSc patients compared with controls [1.730 (0.522) vs 1.525 (0.222) mm; P = 0.039]. The presence of dysphagia in SSc patients was related to a thicker duodenal wall [1.888 (0.578) mm; P = 0.004]. Patients with Medsger Disease Severity 1 had a significantly thicker duodenal wall [1.742 (0.550) vs 1.693 (0.467) mm; P = 0.002], submucosa [0.434 (0.129) vs 0.405 (0.119) mm; P = 0.02] and muscularis [0.691 (0.297) vs 0.633 (0.157) mm; P = 0.003] compared with patients with Medsger Disease Severity 0. Again, no significant differences of the duodenal wall and its layers could be observed between dcSSc and lcSSc patients (Table 2), and between patients with short vs long disease duration (data not shown). Patients with RP not requiring vasodilators tend to have a thinner duodenal wall and layers compared with patients with digital tip ulcers, though this did not reach statistical significance (data not shown).

Discussion

Despite a small sample size, this study provides evidence that involvement of the upper GI tract in patients with SSc leads to a thickening of the mucosa, submucosa and muscularis, which is based on the accumulation of the extracellular matrix.

We hypothesized that the functional abnormality of the GI tract would be associated with a characteristic qualitative or quantitative morphology of the GI wall resulting from fibrosis. With regard to the histological findings [5, 8–10], we expected a thickening of the GI wall in SSc patients compared with controls.

In order to obtain optimal results, the definition of the control group was crucial. Aiming to establish a control group that would resemble the SSc cohort in most aspects, we included persons with a comparable degree of unspecific but without severe GI symptoms and without inflammation-, malignancy- or vascular-mediated thickening of the GI wall.

The study demonstrates that in comparison with control patients, SSc patients presented with thicker upper GI walls. Significant differences were observed for the oesophageal, the antral and the duodenal wall. Dysphagia, the leading clinical symptom for GI involvement, predisposed to even thicker walls. In contrast, the walls of the oesophagus, antrum and duodenum were within the range of controls in SSc patients without dysphagia. Patients with GI Medsger Disease Severity 1 also had significantly thicker walls compared with those with Medsger Disease Severity 0. GI involvement in SSc obviously is associated with a thickening of the upper GI walls.

Earlier studies had described normal histopathological findings in oesophageal segments that showed reduced contractility. This might have been related to methodical problems of the histopathological examination [11–13]. Meanwhile, there are several observations demonstrating functional abnormalities [5, 14] as an altered function of enteric neurons and a loss of smooth muscle with progressive fibrosis [2], and substantial evidence has accumulated that generalized GI dysmotility can be mainly attributed to abnormal collagen deposition and fibrosis during disease progression [15, 16]. The thickening of the GI walls is in accordance with these observations.

The most commonly involved GI organ is the oesophagus [12]. Our results are also in accordance with this observation. The deposition of collagen appears to be located within the submucosa and the muscularis since these layers are significantly thicker. This observation is consistent with the results of an earlier study [5], demonstrating qualitative differences of the oesophageal wall between SSc patients and healthy subjects detected by ultrasound.

Several clinical studies [2, 13–15] and a few morphological analyses [5, 7, 16] have been performed on the oesophagus in SSc patients, but there are only a few studies addressing the stomachs of these patients. Two studies describe the abnormal function of the stomachs of SSc patients [14, 15]. One study has examined the muscular coat of the gastric wall in a case of SSc [17]. In addition, delayed gastric emptying and abnormal postcibal antral motility have been documented [18, 19]. Two studies focusing on the same patients analysed morphological features: Type I and III collagen in the lamina propria has been observed to increase towards the lamina muscularis mucosae, which showed features of atrophy with wide areas of focal fibrosis surrounding smooth muscle cells [9]. Furthermore, a prominent T-cell infiltration with a significantly increased CD4⁺/CD8⁺ cell ratio was detected in SSc patients compared with control patients. The expression of fibrogenic cytokines including TGF-β, connective tissue growth factor, ET-1 and α-smooth muscle actin was increased in SSc patients compared with control patients [10]. Our results are in accordance with these findings. Patients with SSc presented with significantly thicker antral walls, which was most likely related to a significant thickening of the submucosa and muscularis.

The significant thickening of the stomach wall was only observed in the antrum but not in the corpus. This may be due to the shape of the corpus since the corpus is more distensible than the antrum, leading to a greater variation of wall thickness depending on the state of filling.

Only a few studies [9, 10, 14, 15, 17–19] have been performed on the stomachs of SSc patients, but there is even less literature concerning their duodenum [19, 20]. One study described a significant increase in myoelectrical activity after water instillation in the duodenum and an absence of myoelectrical response after direct hormonal stimulation with secretin or pentagastrin [14, 20]. Another study reported a reduced proximal small bowel postcibal motility and uncoordinated fasting or postcibal hypermotility suggestive of a neuropathy [19]. Our study demonstrates a significant thickening of the duodenal wall. This may be caused by abnormal collagen deposition and fibrosis by the same pathogenetic pathway as in the oesophagus and the stomach.

The thickness of the upper GI wall and its layers does not differ between the dcSSc and lcSSc patients. This is in accordance with the fact that the subclassification is only based on a clinical discrimination. It does not reflect the extent of the GI involvement.

In conclusion, our data demonstrate that there is a significant thickening of the walls of the upper GI tract in patients with SSc with the submucosa and muscularis predominantly affected. Clinical signs as dysphagia and amotility are associated with even thicker walls. These results strengthen the hypothesis that increased deposition of different collagen types is an important aspect in the pathogenesis of GI involvement in SSc. Fibrosis and muscle atrophy may follow as proposed in the literature [9, 10], and lead to the functional disorders and clinical symptoms commonly observed in SSc patients.

Disclosure statement: U.M.-L. is supported by a grant from the German Ministry of Education and Research (BMBF). All other authors have declared no conflicts of interest.