Association of Constipation with risk of end-stage renal disease in patients with chronic kidney disease

Chronic kidney disease (CKD) is a global health issue. If CKD couldn’t be adequately controlled, it would progress to end-stage renal disease (ESRD), which requires costly intervention of renal replacement therapy and will be a great burden to family and government [1]. There are many risk factors for CKD, including diabetes mellitus, hypertension, analgesics, herbs, kidney stones, and infection, etc. [2‐8]. However, even under strict control, CKD may still progress. Some other unknown risk factors may be ignored.

In the near decades, human gut was found to harbor 10¹¹ to 10¹² microbiota which influence the nutrition, metabolism, and immune function of the host and has been recognized as an endogenous organ inside the body [9‐12]. In healthy people, the gut microbiota interact harmoniously with the host in mutualistic relationship, so called symbiosis. However, many chronic diseases will disturb the gut microbiome and turn symbiosis to dysbiosis, which was related to some complications and the worsening of the chronic disease [13]. The interaction between kidney and gut, so called kidney-gut axis, was one of the concerned [14]. In the “kidney-to-gut” side, CKD is associated with fluid overload with intestinal wall edema, accumulation of uremic toxin, decreased consumption of dietary fibers, oral use of phosphate binders and/or iron, and so on. These factors all contribute to intestinal dysbiosis [12, 15]. Constipation is the most common gut abnormality in CKD patients. In the opposite “gut-to-kidney” side, gut microbiota also produce some uremia retention molecules, like indoxyl sulfate and p-cresyl sulfate, which are fully excreted by the normal kidney but would accumulate as kidney fails [16]. The increased levels of these metabolites are associated with chronic inflammation, increased cardiovascular mortality and the progression of CKD [17‐19]. The interaction between kidney and gut would be a vicious cycle that worsen the kidney further. Constipation might be an indicator of gut dysbiosis in CKD patients and suggested an ongoing vicious cycle. However, the constipation in CKD patients was usually ignored and not even thought as a risk factor of the progression of CKD.

Whether the CKD patients with constipation had poor profiles of gut microbiome inside was not known yet. We hypothesized CKD patients with constipation may suffer from rapid progression of kidney disease for the negative effects of poor profiles of gut microbiome. The current study tried to explore the relationship between CKD progression and constipation by analyzing a nationwide database.

The demographics, socioeconomic status, concomitant diseases and related medications between the two groups with and without constipation were presented in Table 1. Among the 7818 newly diagnosed cases of CKD, 3909, or 50%, had constipation and the other 50% had no constipation during the follow-up time. Table 1 presented the patients’ characteristics in two years before index date. Data for demographics, socioeconomic status, concomitant diseases and related medications showed no difference in distribution between groups.

In follow-up to the end of 2013, 371 patients (9.5%) experienced ESRD in those with constipation as compared to 182 (4.7%) in those without constipation, respectively (Table 2). Over the 13-year period, the incidences of ESRD per 1000 person-years were 22.9 for constipation group and 12.2 for non-constipation group. Figure 2 showed that there was statistically significant difference for cumulative incidences for ESRD between the patients with and without constipation (P < 0.0001). Cox proportional hazard models with a time-dependent variable revealed an adjusted HR of 1.90 (95% CI, 1.60–2.27) in the CKD patients with constipation (Table 2). Compared to the CKD patients without constipation, adjusted HR for the CKD patients with laxatives < 33, 33–197 and ≥ 198 days per year were 0.45 (0.31–0.63), 1.85 (1.47–2.31) and 4.41 (3.61–5.39) respectively. Adjusted HR for the CKD patients with laxatives ≤ 1 type and > 1 type during the study period were 2.08 (1.76–2.50) and 1.53 (1.19–1.95), respectively, compared to the non-constipation group.

To investigate the impact of kidney-gut axis, our study compared the difference of occurrence of ESRD between CKD patients with newly developed constipation and without constipation. The baseline of demographic and socioeconomic status was matched by propensity scores. The well-known comorbidities related to the progression of CKD and the nephrotoxic medications were also included and matched. Furthermore, the concomitant diseases and medications that might cause constipation were also added and matched. The 2 groups, CKD with constipation and CKD without constipation, were fairly matched and then followed longitudinally from 2000 to 2013. The ESRD events was 371, or 22.9 per 1000 person-year, in CKD with constipation and 182, or 4.7 per 1000 person-year, in CKD without constipation respectively. The cumulative incidence for ESRD in CKD with constipation was significantly higher than those without. The adjusted HR for CKD with constipation was 1.90 (95% CI, 1.60–2.27). The result showed de novo constipation in CKD patients increased the risk of the development of ESRD. To clarify the relationship between CKD and constipation, we further analyzed the incidence of ESRD in constipation patients with different duration of laxatives use per year and different numbers of types of laxatives. Compared to the CKD patients without constipation, adjusted HR for the CKD patients with laxatives < 33, 33–197 and ≥ 198 days per year were 0.45 (0.31–0.63), 1.85 (1.47–2.31) and 4.41 (3.61–5.39) respectively. For mild constipation, especially those who nearly did not need laxatives (laxatives < 33 days), the HR (0.45) of ESRD did not increase. However, for obvious constipation, with laxative use more than 33 days per year, the HR (1.85 for 33–197 and 4.41 for ≥198) of ESRD significant higher than non-constipation group. For those with even longer duration of laxatives use, more than 198 days per year, the HR (4.41) was higher than those with laxative use between 33 and 197 days per year (1.85). The result showed the mild constipation, with nearly no laxatives needed, would not increase the risk of ESRD; however, the obvious constipation, with laxatives use more than one month per year, would increase the risk of ESRD. Furthermore, more severe constipation, with laxatives use more than 6 months per year, would bring even higher risk of ESRD. There seemed to be a dose-effect relationship between CKD and constipation, which consolidated our hypothesis. The adjusted HR for the CKD patients with laxatives less than or equal to one type (≤ 1 type) and more than one type (> 1 type) during the study period were 2.08 (1.76–2.50) and 1.53 (1.19–1.95), respectively, compared to the non-constipation group. Both ≤ 1 type and > 1 type laxatives showed significantly increased HR, but the > 1 type did not have higher HR than ≤ 1 type. Combining the analysis above, the duration of laxative use should account more in the risk of ESRD than the different numbers of types of laxatives use.

The main limitation of this study is that we cannot rule out the possibility of residual confounding by unmeasured factors such as lifestyle and behavioral factors such as diet, smoking, alcohol drinking and exercise. Otherwise, we tried to reduce the confounders related to the progression of CKD and the development of constipation. To reduce the confounders related to the progression of CKD, several associated comorbidities, including acute coronary syndrome, diabetes, hypertension, hyperlipidemia, COPD and cerebrovascular disease, were added to the propensity score-matching models (Table 1). Previous use of NSAIDs and analgesic drugs other than NSAIDs were also added to matching. To reduce the confounders related to the development of constipation, the concomitant diseases and related medications which may cause constipation were also added into the matching. We tried to find if the de novo constipation per se increased the risk of ESRD in CKD patients.

Constipation status and severity was reported as a risk factor of incident CKD and ESRD [24]. It suggested healthy people needed to monitor their defecation status to maintain kidney health. Our study focused on the CKD people and suggested they should maintain good bowel habit to avoid de novo constipation which might precipitate the CKD progression. Moreover, even the constipation could be controlled by the laxatives, but the erroneous gut microbiome, that caused the negative effect to kidney via kidney-gut axis, might not be changed. Therefore, some other interventions, like the use of probiotics and prebioitics [25‐27], smart bacteriae [28], and high-fiber diet [29], or the application of toxin-absorbing agents [30, 31] should be considered to correct this error. It is also relevant to define the specific microbiome profiles associated with CKD and ESRD to give the precise therapy. Future studies should explore the profiles of microbiome in CKD and the interaction between them.

Our longitudinal nationwide study showed that de novo constipation in CKD patients was a risk factor in progression to ESRD. The obvious constipation, with laxatives needed, increased the risk. The more severe constipation, with the longer duration of laxatives needed, increased the risk further. The patency issue of bowel should not be ignored in CKD patients. Dietary modification and some other interventions should be tried to avoid constipation. Avoidance or even shortened intervals of constipation might be helpful to lessen the progression of CKD.