Autonomic neuropathic changes are frequent in patients with diabetes and also in those suffering from chronic kidney disease [3, 10, 11]. Especially cardiac autonomic neuropathy (CAN) may result in severe consequences such as sudden death, myocardial ischemia and overall increased mortality [2, 4]. Using COMPASS 31, a new self-assessment test, that is well defined and shows good reliability, we could document much higher AN-scores in dialysis patients compared with controls (Fig. 2). As there is a strong correlation between AN-score and diabetes duration (Fig. 3a) and as in the dialysis patients’ group diabetes duration is longer than in controls (23,4 vs. 12,5) this potential effect should be taken into account as a possible limitation. However, even after dichotomization of diabetes duration for dialysis patients’ group (AN-score lower or higher as the median of 16 years), there was still a significant difference for each of the groups compared with controls (group with diabetes duration < 16 years: p < 0,05; mean for AN-score 25,8 vs. 8,4; group with diabetes duration > 16 years; p < 0,01; mean for AN-score 29,7 vs. 15,3). Increased AN-scores were found not only for cardiac autonomic alterations, which mostly have been in the focus of preceding investigations, but also for neuropathic disorders of other affected organ systems (secretomotor, gastrointestinal, bladder, pupillomotor subtypes). To our knowledge this the first investigation in dialysis patients involving comprehensively these subtypes of neuropathic changes besides signs of CAN by a self assessment test.

Incidence and progression of AN in dialysis patients at least partially are due to uremic toxins [12, 13]. Besides uremic toxins also advanced glycation endproducts (AGEs) play an important role in the genesis of autonomic neuropathy, especially in patients with diabetes disease [14]. Moreover various reactive oxygen metabolites and oxidative stress contribute to the development of AN, as well as disturbances in the polyol and other pathways [3]. Abovementioned substances partially unfold their neurotoxic effects as a consequence of their insufficient removal by dialysis therapy. If dialysis therapy is able to improve AN is not generally proven. Some authors report partial regression of AN after initiation of dialysis therapy while others could not confirm a clear correlation between time period after starting dialysis therapy and extent of autonomic neuropathic symptoms [15, 16]. Patients with pronounced autonomic neuropathic disorders are at high risk for cardiovascular events and death. High AN-scores eventually could enable us to treat these patients more intensively. With respect to renal replacement therapy this could possibly result in hemodiafiltration treatment instead of hemodialysis treatment.

In the multimorbid collective of dialysis patients many of them are suffering not only from abovementioned metabolically induced changes but also from ongoing vascular damage with long-lasting diabetes. This may further aggravate AN and additionally explain the highly significant correlation between AN-score and overall duration of diabetes disease in our study cohort.

Current HbA1c was not associated with AN-scores in our collective. It can be hypothesized that not all dialysis patients had good monitoring of diabetes disease all the time before starting renal replacement therapy. On the other hand, it can be assumed in some cases that glycemic control was even improved with the initiation of dialysis therapy, at a moment where longlasting negative effects on autonomic nerve function had resulted in severe AN. Low HbA1c in dialysis patients partially might be attributed to compromised appetite, reduced body weight and thus improved insulin resistance (lower body weight in this group). Furthermore dialysis patients are usually seen three times a week and thus have a frequent and tight relation with their nephrologist. The effect of improved medical care in improving glycemic control in these patients should not be underestimated.

The missing correlation of AN-score with BMI may be explained by the fact that most dialysis patients are suffering from long existing neuropathic changes, while on the other hand during dialysis therapy loss of appetite, malnutrition and reduced general condition are increasing. Thus, potential influences of BMI on AN-score may change with time and possible correlation may disappear.

In some studies height was correlated with severity of AN and attributed to increasing length of nerve fibers with greater body height. Other authors, however, could not find such an association in a population study, and are in accordance with our results [17, 18]. Age and AN-score are positively correlated, as shown in Fig. 3b. Statistical significance, however, was narrowly lost when analyzing both control group and dialysis patient group separately (Fig. 3c and d).

It is well known that chronic inflammation is enhanced in chronic kidney disease and especially in dialysis patients [19,20,21,22]. This could be confirmed in our study and we additionally found a positive correlation with AN-score. To what extent this association of CRP and AN-score is due to inflammation, vascular changes or mediators of other pathways can not be answered by our study.

A limitation of our study might be, that in dialysis patients the anuric state may impact the subdomain “bladder”. As anuria was present in only 14% of dialysis patients and due to the fact that there was still diuresis of more than 500 ml in 55% and more than 1000 ml in 31% of the dialysis patients, we think this is in total acceptable. Also the AN-score for bladder is weighted only with a factor of 0.33. Moreover the self assessment test was done before starting of the daily dialysis treatment. The questionaire includes time-dependency of dry eye problems and of sweating in general (have you recognized an increased frequency of dry eyes or mouth in the past?). This would detect developing of polyneuropathic symptoms in time course. Nevertheless there could be effects of actual hydration and dry weight changes, however, more likely in orthostatic and vasomotor problems. This should be taken into account but remains a difficulty in most studies in dialysis patients. Generally also antihypertensive treatment may influence changes in these domains.

Future cross sectional and longitudinal studies with higher case numbers might help to use results from the COMPASS 31 tool for better detection of autonomic neuropathic changes in the variety of affected organ systems.