Distal Gastrectomy with Billroth II Reconstruction is Associated with Oralization of Gut Microbiome and Intestinal Inflammation: A Proof-of-Concept Study

Fourteen patients were included in the SGB2 group, and eight participants in the control group. The baseline clinicopathological characteristics are presented in Table 1. Six (75.0%) controls were spouses, and two (25.0%) were children. Participants in the control group were younger than the patients and predominantly female. The median time from surgery to enrollment was 45 (Q1;Q3: 26;63) months, while the minimum and maximum times were 6 and 101 months, respectively. Three (21.4%) patients received adjuvant chemotherapy following surgery, although in all cases, at least 12 months prior to the enrollment. None of the patients had disease recurrence at time of enrollment (Fig. 1).

We investigated the alteration in the fecal microbiome of patients after SGB2. Our results clearly show the impact of SGB2 on the general gut microbiome composition, with decreased alpha diversity by Shannon index after SGB2 and significant differences in beta diversity between patients and healthy controls as well as taxonomic composition. Taxon comparisons revealed that approximately half of the genera with altered abundance have been linked to PPI therapy in previous studies. PPI intake increases the gastric pH from the physiological level of approximately 2.0 to over 6.0,21 considerably higher than pH 4, which is considered to be the threshold value for powerful bactericidal effect.22 Similar to PPI intake, SGB2 causes permanent increase of the gastric pH to values above 6.0.23 Therefore, our findings can be explained by the comparable loss of gastric barrier function after SGB2 and by PPI use. Vice versa, our results support the notion that PPI-induced microbiome changes are caused by acid suppression and are most likely not due to direct drug-induced effects on microbes.

The steep increase in Escherichia–Shigella was the most prominent difference between the microbiome of SGB2 patients and that of healthy controls. Escherichia is a common protagonist in small intestinal bacterial overgrowth (SIBO),24 which occurs in the majority of patients after gastrectomy and is associated with intestinal and postprandial symptoms.25 A similar observation was made in children after PPI therapy.26 Although members of the genus Escherichia‐Shigella are not sensitive to pH variations in their environment, these seem to profit from the altered milieu, since these were also found to be increased in the general population after PPI intake.8^,27 The observed increase in Enterococcus, a bacterium that is also often involved in SIBO, however, is directly attributable to the increased gastric pH. In a model of gastric barrier dysfunction, both genetic and pharmaceutical blockage of acid secretion in the stomach resulted in increased survival of orally gavaged Enterococcus.15 Moreover, after SGB2, patients showed a significant increase in Streptococcus. Streptococcus is a prevalent bacterial taxon in the oral cavity and the most commonly described bacterium in PPI-induced dysbiosis.8^,9^,11^–13 This was recently linked to intestinal inflammation and gut permeability in cirrhosis patients.14 In the present study, we showed that Streptococcus is also associated with intestinal inflammation in patients after SGB2. Together with other oral bacteria (Veillonella, Oribacterium, and Mogibacterium), the observed increase in Streptococcus abundance supports the hypothesis of oralization after gastric acid barrier disruption, also in patients after SGB2. Furthermore, several beneficial commensals were decreased in the microbiome of SGB2 patients. The loss of these commensals correlated with the increase in calprotectin levels in stool. Especially the diminished abundance of Faecalibacterium, Subdoligranulum, and members of the Ruminococcaceae and Lachnospiraceae family again is similar to PPI dysbiosis.9^,11^,12^,14

Besides the important pathophysiological information, our study may also have clinical implications for patients after SGB2. Chronic intestinal inflammation after SGB2 plays an important role in the patients’ health and quality of life. Although overall quality of life scores show an immediate deterioration after surgery followed by an increase to approximately normal levels within the first year, gastrointestinal symptoms remain a significant issue long after SG.28^–30 In the present study, calprotectin levels were markedly increased in SGB2 patients and strongly associated with the presence of Streptococcus in the stool. A very similar pattern can be found in patients with long-term PPI use, in whom increased calprotectin levels and associations between oralization and inflammation have been described in previous reports.14^,31^,32 Chronic intestinal inflammation has been described in the pathogenesis of chronic diarrhea after SGB2.33 Intermittent or permanent chronic diarrhea is one of the most common problems in long-term survivors after gastrectomy,28^,34^,35 present in about 40% of patients.36 In the present study, approximately 54% of patients also suffered from diarrhea and showed higher calprotectin levels on average than patients without diarrhea, although this observation did not reach statistical significance, and validation in bigger studies is warranted. In patients with diarrhea, Ruminococcus 1 was depleted, and Mogibacterium was overrepresented. Ruminococcus 1 is a ubiquitous genus in the human microbiome that has the ability to degrade complex carbohydrates and provide nutrients for other commensals.37Ruminococcus species have been associated with a stable human microbiome in previous reports,38 and decreased abundance was associated with diarrhea in a porcine animal model.39Mogibacterium was found to be increased in Crohn’s disease and colorectal cancer patients.40^,41 Other common gastrointestinal symptoms were abdominal discomfort and bloating. Both symptoms were associated with a decrease of Agathobacter. Agathobacter are butyrate producers who live in symbiosis with Bifidobacteria, giving them access to acetate as a substrate for butyrate production.42 Moreover, an increased abundance of Holdemanella was observed in patients with abdominal discomfort. Comprehensive studies on Holdemanella on human health are lacking, however, their taxonomic family Erysipelotrichiaceae contains highly immunogenic species and is associated with proinflammatory conditions.43 Interestingly, patients who reported bloating also showed a reduced abundance of Streptococcus. Streptococcus is the foremost genus in PPI-associated dysbiosis and has been linked to inflammation and gut barrier dysfunction before. However, the genus Streptococcus entails also beneficial species, such as S. salivarius subsp. thermophilus that is utilized in various probiotic products. VSL#3, which contains a Streptococcus species among others, has been shown to reduce bloating in patients with irritable bowel syndrome.44^,45 Similarly, another multispecies probiotic containing S. thermophilus improved self-perceived gastrointestinal wellbeing.46 More in-depth studies are necessary to clarify the role of different Streptococcus species in gastrointestinal health and disease. Nevertheless, the associations between gastrointestinal symptoms and the microbiome in SGB2 patients highlight the importance of comprehensive studies in this field to improve patients’ postoperative outcomes and wellbeing.

Acid-unrelated changes in the microbiome of SGB2 patients include an increase of Oxalobacter abundance. Oxalobacter is an oxalate-metabolizing commensal that increases the colonic excretion of oxalate, which in turn, reduces the strain of calcium oxalate on the kidney.47 In the present study, Oxalobacter was exclusively found in patients after SGB2 and was absent in healthy controls. Although clinical trials that utilized Oxalobacter as a probiotic in patients with primary hyperoxaluria were unsuccessful,48 the natural occurrence of Oxalobacter after SGB2 might be a beneficial adaptation to the altered gastrointestinal physiology after SGB2.

The microbiome faces a variety of influencing factors, such as diet, gender, and age of the patient, that also need to be considered in cohort studies. By selecting in-house relatives as controls, we minimized the diet-related impact on gut microbiome composition as similar microbiome of individuals who share a household has already been shown previously,49^,50 but we had to accept an age and gender bias. Our multivatriate analysis showed that the impact of age and gender was overshadowed by the strong influence of SGB2 on the microbiome composition. This was not unexpected since the age difference between the groups was rather small, and the changes in the microbiome after SGB2 such as the steep Enterococcus increase were more dominant compared with changes due to age. However, comparisons are hard to draw, since data on the aging microbiome are limited, and the findings are inconsistent.51^,52 Gender-related differences in microbiome composition have been previously described in health and disease.53^–55 Natural male predominance in the gastric cancer group and the expected female predominance in our control group might hinder the detection of gender-related differences further. Chemotherapy may also have an impact on gut microbiome composition. Dysbiosis has been described in the short term after chemotherapy application and linked to mucositis and impaired capability to resist pathogen colonization.56^,57 However, there is a lack of data supporting whether dysbiosis persists in the long term, while this is still under investigation in an ongoing study.58 Chemotherapy may have some long-lasting slight impact on the gut microbiome composition, potentially similar to long-lasting imprint described in healthy adults after exposure to short-term broad-spectrum antibiotics.59 Therefore, in our present study, we could not rule out history of chemotherapy as a potential cofounder affecting microbiome, and excluded patients who received chemotherapy within the past 12 months.

Our results are in stark contrast to previously published sequencing data in patients with SG and B2 or RY reconstruction.60 In said study, the genera Oxlobacter, Veillonella, Streptococcus, Escherichia, Shigella, and Oribacterium among others were attributed to the control groups, while these were a crucial part of the microbiome alteration after SGB2 in the present study. Although the previous study had a rather big sample size, healthy controls were insufficiently characterized, and the use of medication was not analyzed as a potential confounder, which might lead to misinterpretation of the results. As we showed in our study, changes after gastrectomy can mimic drug-induced changes in the microbiome and, therefore, obscure the effect of the surgery. Especially, gastric pH-associated changes might be vulnerable to uncharacterized drug use in the control groups since PPI use is among the most dominant confounders in microbiome analysis in the general population.61 Large well-characterized cohorts are needed to fully elucidate this topic.

Our proof-of-concept study has several limitations. First is the relatively small sample size of the study. To prove the concept of increased gastric pH-related changes in the microbiome, the cross-sectional design of the present study was sufficient, although this is lacking data to show microbiome composition changes pre- and post-SGB2. Even with the relatively small but homogenous cohort and well-selected controls of this study, we were able to clearly confirm our hypothesis and show that SGB2 is associated with changes in the gut microbiome that can be attributed to the increased gastric pH. Second, our study investigated the fecal microbiome composition only in patients who underwent SG with B2 reconstruction. Therefore, it remains unclear whether other types of anastomosis, such as B1 or RY, might have the same impact on the gut microbiome. Future studies including all types of anastomosis will be important for generalization of our findings. However, since B1 gastroduodenal anastomosis is a common technique, especially in Asian countries,62 and RNY is the preferred method in Western countries,63 these studies might require prospective multicenter studies on an international scale. However, the same increase of gastric pH to the level above 6 has been reported after SG irrespective of B1 or B2 anastomosis; 23 therefore, it seems likely that the oralization of the gut microbiome phenomena would be attributable to the SG itself, irrespective of the reconstructive method.

Our study shows that SGB2 is associated with microbiome oralization and intestinal inflammation. These findings prove that an increase in gastric pH irrespective of the reason for this increase is associated with typical microbiome changes. These novel findings may open gut microbiome as a new target for therapy to improve quality of life and general patient health in long-term survivors after SGB2.