Shiftwork, functional bowel symptoms, and the microbiome

There are about 15 million Americans working full-time on evening, night, or rotating shifts, or other irregular employer-arranged schedules; 4.7% on evening shifts, 3.2% on night shifts, 3.1% on irregular schedules, and 2.5% on rotating shifts (). Night shift work is associated with increased mortality, higher risk of cardiovascular disease, cancer, diabetes, hypertension, chronic fatigue, sleep problems and higher body weight (;;;;;). Night and rotating shift workers also report a higher prevalence of Irritable Bowel Syndrome (IBS), abdominal pain, constipation and diarrhea than do day shift workers (;;;). In fact, between 48% and 82% of those working rotating or night shifts report abdominal pain, constipation, diarrhea and other symptoms of functional bowel disorders (;). [United States Department of Labor Bureau of Labor Statistics, 2005] [Gu et al., 2015] [Jia et al., 2013] [Myers et al., 2015] [Pan et al., 2011] [Rajaratnam et al., 2011] [Vyas et al., 2012] [Caruso, Lusk & Gillespie, 2004] [Knutsson & Boggild, 2010] [Nojkov et al., 2010] [Wells, Roth & Chande, 2012] [Nojkov et al., 2010] [Saberi & Moravveji, 2010]

The basis for this high prevalence of functional bowel disorders, including IBS, among shift workers is unknown. However, some studies suggest that inappropriate nutrition or irregularity in the timing of meals (;), and psychological disorders () may contribute to the high prevalence of functional bowel symptoms among workers on rotating or night shifts. [Bilski, 2006] [Lowen et al., 2010] [Zhen, Ann & Yu, 2006]

Other studies strongly suggest that sleep deprivation or sleep disturbances are associated with the presence and severity of functional bowel symptoms reported by resident physicians and nurses (;;). Moreover, animal studies suggest that circadian disruption, similar to that in shift workers, contributes to the development of gastrointestinal (GI) complaints among shift workers by altering the composition and normal diurnal rhythmicity of the resident intestinal microbes (). Interestingly, gut microbiota community composition and diversity are malleable and sensitive to changes in diet and other environmental factors (), including activity and sleep patterns. For example, Thaiss and colleagues () “jet lagged” a group of mice by subjecting them to an 8-hour advance for three days before allowing them to revert to their usual schedule for three more days, then subjecting them to another 8 h advance for three days. Mice exposed to 4 weeks of this schedule lost their usual pattern of physical activity, and consumed food at irregular intervals. Significantly, this environmentally induced disruption of daily activity patterns (jet lag schedule) was associated with a loss of diurnal rhythmicity of microbiota composition in mice. [Jarrett et al., 2000] [Saberi & Moravveji, 2010] [Wells, Roth & Chande, 2012] [De Bacquer et al., 2009] [Voigt et al., 2014] [Thaiss et al., 2014]

Thaiss and colleaguesalso found similar changes in human microbiota composition in two volunteers who flew from the US to Israel (an 8-10 h advance). Samples collected at baseline (one day pre-flight), during jet lag (one day after landing), and during recovery (2 weeks after landing) showed rapid changes in the composition of the microbiota. During jet lag (the first 24-hours after landing), there was a higher relative representation of Firmicutes, which reversed upon recovery from jet lag (2 weeks later). It is unknown it these changes would occur with the chronic circadian disruption experienced by night shift workers. Although some studies have demonstrated no differences in composition of the fecal microbiome when samples from lean and obese individuals were compared (;), other studies in humans have demonstrated that Firmicutes are associated with a higher propensity for obesity and metabolic disease (;), conditions that are more common in night and rotating shift workers (;). [(2014)] [Ley et al., 2006] [Turnbaugh et al., 2009] [Finucane et al., 2014] [Ley et al., 2006] [De Bacquer et al., 2009] [Suwazono et al., 2008]

Finally, multiple studies have linked reduced microbial diversity and richness in microbial communities to IBS symptoms. For example, Krogius-Kurikka and colleagues () reported that fecal samples from patients with diarrhea-predominant IBS were enriched with Proteobacteria and Firmicutes but had reduced Actinobacteria and Bacteroidetes compared to healthy controls. Other studies (;) have shown an increase in the Firmicutes-to-Bacteriodetes-ratio, a decrease in some types of Firmicutes families (Lactobacilli, Faecalibacterium) and the Actinobacteria population (Bifidobacteria, Collinsella), and an increase in some Firmicutes families (Veillonella, Streptococci, and Ruminococcus spp.) and in Proteobacteria (Enterobacteriaceae spp.). In addition, low microbial richness, an absence of Methanobacteriales, and enrichment with Bacteroides enterotypes are associated with more severe IBS symptoms (Tap 2017). Not only is the composition of the fecal microbiota altered in patients with IBS, these imbalances in the microbial community or dysbiosis, occur more frequently in patients with IBS compared to healthy individuals. Reduced diversity was observed in nearly three-fourths of the IBS patients studied by Casén and colleaguescompared to 16% in normal individuals (;). [Krogius-Kurikka et al., 2009] [Bhattarai, Pedrogo & Kashyap, 2017] [Salonen, de Vos & Palva, 2010] [(2015)] [Collins, 2014] [Jeffery et al., 2012]

While these studies suggest that shiftwork alters gut microbiota and that alterations in gut microbiota are common in patients with IBS, they do not demonstrate whether these alterations are associated with the somatic symptoms experienced by many rotating and night shift workers. Therefore, the present study is designed as the initial step in determining if there are differences in (1) composition and diversity of the microbiome of night shift workers compared to day shift workers; and (2) the composition and diversity of microbiome among night shift workers experiencing functional bowel symptoms (e.g., bloating, lower abdominal pain, constipation and diarrhea) compared to night shift workers not experiencing functional bowel symptoms.

The PCoA plot inshows that samples obtained from the same participants tend to cluster together and samples from the same time on sample participants tend to cluster even more closely. This observation is consistent with the notion that microbiome samples are homogeneous within subjects and heterogenous across subjects, and thus confirms good quality of our data. Supplemental text (phyloseq analysis) contains plots showing the composition of the samples at taxonomic different levels Fig. 1

As shown in, there were no significant differences in Chao1 (= 0.411 by the LDM) and Shannon (= 0.242) between day and night shift nurses. Nor were there differences in Bray-Curtis (= 0.476 by PERMANOVA) and Jaccard distances (= 0.625) by shift type (). Nor were there differences in relative abundance (= 0.489 by the LDM) and presence-absence (= 0.824) data across all genera. Fig. 2 Fig. 3 p p p p p p

shows an increase of the log Chao1 index from the beginning to the end of the shift for day-shift workers while a decrease for night-shift workers, and this difference of change is significant (interaction= 0.034 by the LDM); so does the Shannon index (= 0.008). However, the change among the pooled workers was not significantly different (main effect of time= 0.473 for Chao1 and= 0.236 for Shannon) possibly due to the cancellation of effects with opposite directions. In terms of beta diversity, the change of the beta diversity metric from the beginning to the end of the shift among the pooled workers was marginally significant or significant (main effect of time= 0.056 by PERMANOVA for Bray-Curtis and 0.014 for Jaccard). Marginally significant and significant findings were also noted by the LDM (main effect of time= 0.035 and 0.063 based on relative abundance and presence-absence data, respectively). However, there is not enough evidence to confirm that the change was significantly different between day shift and night shift workers possibly due to the small sample size, although there was suggestive evidence (interaction= 0.192 and 0.118 by PERMANOVA based on Bray-Curtis and Jaccard, respectively;= 0.320 and 0.134 by the LDM based on the relative abundance and presence-absence data). In addition, the LDM revealed seven ASVs to be differentially abundant between the beginning and the end of the shift; ASV_455(S5-A14a, more abundant at the beginning), ASV_2527(Ruminoccaceae_NK4A214_group, more abundant at the end), ASV_1304(Ruminococcus_1, more abundant at the end), ASV_221 (Mobiluncus, more abundant at the beginning), ASV_7(Campylobacter), ASV_130(Alistipes, more abundant at the end), and ASV_62(Agathobacter, more abundant at the end). Figure 4 p p p p p p p p

Finally, as shown inthere were no significant difference in alpha diversity when comparing participants without and with IBS symptoms (= 0.849 for Chao 1,= 0.484 for Shannon, by the LDM method). Although there were no differences in beta diversity (= 0.206 for Bray-Curtis and= 0.213 for Jaccard) by whether or not the participant had symptoms of IBS), there were some significant differences based on the LDM results. Specifically, three ASVs, ASV_1160 (Flavonifractor), ASV_1134 (Oscillibacter) and ASV_2379 (Ruminiclostridium_9), were detected to be differentially abundant. Three ASVs, ASV_1160 (Flavonifractor), ASV_2379 (Ruminiclostridium_9) and ASV_47 (Escherichia/Shigella), were detected to be significantly more likely to be present in participants reporting IBS symptoms. Fig. 5 p p p p

The findings of this pilot study suggest that there are no differences in the richness and diversity of species when samples from nurses working day and night shifts were compared. There were however some statistically significant changes in both alpha and beta diversity metrics when specimens collected at the beginning and end of the shifts were compared and there was also some evidence that the changes were different for day shift and night shift workers, with increased alpha diversity noted at the end of the day shift and decreased alpha diversity noted at the end of the night shift. Seven ASVs were found to be differentially abundant between the beginning and end of the shifts for the entire sample. In addition, there were three ASVs to be differentially abundant in participants reporting IBS symptoms.

Studies comparing the effects of shift work on the gut microbiome are limited and somewhat contradictory. For example, a study of 10 male security guards who worked both day and night shifts found there were no significant differences in in alpha or beta diversity within and across-subject variation for both shifts (). In contrast, slight changes in microbial abundance and diversity were noted when 22 subjects, aged 20–35 years, delayed their sleep period for 2–4 h (). Although there have been studies comparing circadian variation in the gut microbiota in mice () and another describing the results samples collected during multiple time points over several days by two subjects (), our study is the first to compare the richness and diversity of gut microbiota collected from 51 human participants at two different time points in 24 h. [Mortas, Bilici & Karakan, 2020] [Liu et al., 2020] [Thaiss et al., 2014] [Thaiss et al., 2014]

Even though there are numerous studies that have reported increased and/or decreased amounts of various gut bacteria among patients with IBS (;;;;), a recent systematic review found only nine studies that discussed differences in alpha-diversity in patients with IBS compared to normal controls (). Slightly over half of the studies (55.6%) reported a significant decrease in the richness and diversity in patients with IBS (;;;), whereas the remaining four studies (;;;), like our current study, revealed no differences in alpha-diversity compared to healthy controls. A more elaborate study () comparing fecal and mucosa-associated microbiome samples obtained from the sigmoid colon during colonoscopy, not only found no differences in alpha diversity when the two types of samples were compared, but that there were no differences in richness and diversity when samples obtained from healthy controls and patients with IBS or IBS subtypes were compared. [Bhattarai, Pedrogo & Kashyap, 2017] [Casén et al., 2015] [Pittayanon et al., 2019] [Salonen, de Vos & Palva, 2010] [Tap et al., 2017] [Pittayanon et al., 2019] [Carroll et al., 2011] [Carroll et al., 2012] [Liu et al., 2016] [Rangel et al., 2015] [Carroll et al., 2012] [Durban et al., 2012] [Rigsbee et al., 2012] [Tap et al., 2017] [Pozuelo et al., 2015]

Like other studies comparing patients with IBS symptoms to healthy controls, our pilot study found increased Firmicutes (), specifically Flavonifractor, Oscillibacter, and Ruminiclostudium among participants with IBS (). The increased abundance of E. coli/ Shigella possibly reflects the suspected relationship between Shigellosis and IBS (). In contrast,reported that Methanobacteriaceae, Ruminococcaceae, Erysipleotrichacea, and one unknown Clostridiales family were all found to be decreased in patients with IBS-M (IBS-mixed type) and IBS-D (IBS-diarrhea), but not in patients with IBS-C (IBS-constipation). [Chong et al., 2019] [Casén et al., 2015] [Youn et al., 2016] [Pozuelo et al. (2015)]

IBS is estimated to have a world-wide prevalence of 10–15% (;). Prevalence rates of IBS are typically higher among nurses, with rates ranging from 17.4% in China () to 45.2% in Nigeria (). The prevalence rate of IBS among study participants was 35%, quite similar to the prevalence rate reported among nurses at the University of Michigan Medical Center (36.6%) (). However, unlike the study of hospital staff nurses (), and other studies of shift workers (), there were no differences in the prevalence of IBS symptoms among day and night shift nurses in our study. [Canavan, West & Card, 2014] [Sperber et al., 2017] [Liu et al., 2014] [Akere & Akande, 2014] [Nojkov et al., 2010] [Nojkov et al., 2010] [Kim et al., 2013]

This study is limited by a number of factors. First, our study population consisted of a convenience sample of nurses that might not be representative of the larger nursing workforce or the larger population of shift workers. The overall participation rate was relatively low, which raises concerns about how representative the participants were of the total population of nurses who were invited to participate. Additionally, given the focus of the study, nurses who experienced IBS symptoms may have been more likely to participate than those who did not experience IBS symptoms. Finally, the severity of IBS symptoms and quality of life was not assessed, two factors which may have been impacted by gut microbiome diversity and richness.

There were no statistically significant differences in the richness and diversity when samples of the gut microbiome from nurses working day and night shifts were compared. However, when specimens collected at the beginning and ends of the shifts were compared, there were some statistically significant or marginally differences in alpha and beta diversity. Three ASVs were more common in participants reporting IBS symptoms.