Cardiovascular autonomic neuropathy in patients with type 2 diabetes with and without sensorimotor polyneuropathy

Cardiovascular autonomic neuropathy (CAN) in patients with diabetes is associated with poor prognosis. We aimed to assess signs of CAN and autonomic symptoms and to investigate the impact of sensorimotor neuropathy on CAN by examining type 2 diabetes patients with (DPN [distal sensorimotor polyneuropathy]) and without distal sensorimotor polyneuropathy (noDPN) and healthy controls (HC). Secondarily, we aimed to describe the characteristics of patients with CAN.


| INTRODUCTION
Diabetes affects an increasingly large number of people, with an estimated number of 425 million patients suffering from the disease worldwide. One of the most common long-term complications of diabetes is neuropathy, where distal sensorimotor polyneuropathy (DPN) is the most common variant. 1 The symptoms of DPN are either negative somatosensory with reduced sensation to different sensory modalities or positive symptoms with pain and tingling in the lower limbs or a combination thereof. 2 The peripheral nerve damage can also involve the autonomic nervous system. Patients may suffer from length dependent autonomic neuropathy with reduced sweat, color, and skin temperature changes in a glove and sock distribution or a more generalized autonomic neuropathy with multi organ involvement. The generalized autonomic neuropathy can result in cardiovascular, urogenital, gastrointestinal, pupillomotor, and thermoregulatory dysfunctions. [2][3][4] Patients suffering from cardiovascular autonomic neuropathy (CAN) may be asymptomatic but can experience severe and disabling symptoms with, for example, orthostatic intolerance or syncope. Patients with CAN have increased mortality and risk of cardiovascular and renal complications. [5][6][7] Common clinical factors linked to CAN are higher age, diabetes duration, triglycerides, HbA1c, BMI, and presence of retinopathy and nephropathy. 6 Prevalence estimates of CAN in patients with diabetes vary between 20% and 73% in patients with type 2 diabetes and 1%-90% in type 1 diabetes. 8 These studies differ both regarding methods used for diagnosing CAN and in included patient characteristics (e.g., duration of diabetes, age, gender distribution). The diagnostic methodology for the assessment of CAN as stated by The CAN Subcommittee of the Toronto Consensus Panel on Diabetic Neuropathy includes signs, cardiovascular autonomic reflex tests (CARTs) and blood pressure monitoring. 6 The gold standard tests are CARTs, measuring heart rate, and blood pressure responses to provocative physiological maneuvers, including deep breathing, Valsalva maneuver, standing and tilt table test. 9 These methods are usually time-consuming and only performed at specialized centers. A simple tool for the diagnosis of CAN is the handheld Vagus™ device combined with a modified Schellong test. The modified Schellong test includes blood pressure and heart rate measured during rest and during 3 min of passive standing. Vagus™ includes three CARTs corresponding to the gold standard tests described by the Toronto Consensus panel. The device is easy to implement and use in a clinical setting 10 and shares many similarities with the gold standard tests regarding examination of the cardiovagal domain of the autonomic nervous system. The device has compared favorably with a traditional stationary device in a small cohort and shown moderate to high reproducibility. 11,12 The relationship between CAN and DPN is not clear. Prevalence studies are not all supportive of a clear association between DPN and CAN. Some studies have found that CAN is more common in patients with DPN than noDPN in type 1 and type 2 diabetes 13-16 while others have not. 17 None of these studies have used the current definitions of DPN as proposed by the Toronto Diabetic Neuropathy Expert Group. 14 Small nerve fiber damage may affect both peripheral and autonomic small nerve fibers, potentially leading to autonomic dysfunction. 15 Previous studies have investigated small fiber dysfunction in CAN, using simple bedside tools, 15 quantitative sensory testing (QST), 18,19 and corneal confocal microscopy (CCM). 20 While these studies suggest that small fiber dysfunction is characteristic of CAN the association between intra epidermal nerve fiber density (IENFD) and CAN has not been assessed. Assessment of autonomic symptoms has been suggested as a valid approach to diagnose CAN. 6 The first validation study of the COMPASS questionnaire for the diagnosis of CAN found that COMPASS-31 scores were higher both in patients with CAN and DPN, and suggested that COMPASS 31 is an easy and reliable assessment tool for autonomic symptoms of diabetic neuropathy and can be used as a screening tool for CAN. 21 When studying CAN and DPN, there is a lack of systematic appraisal of autonomic symptoms 14,22 and many studies do not assess autonomic symptoms at all. 13,15,17,23,24 We aimed to assess signs of CAN and autonomic symptoms among patients with recently diagnosed type 2 diabetes and wellcharacterized DPN according to the Toronto classification in order to investigate the impact of distal sensorimotor neuropathy on CAN. We hypothesize that these two conditions have a close relationship. Secondarily, we aimed to describe the characteristics of patients who had both CAN and DPN including the association with small fiber dysfunction.

| Study population and study design
The study population in this study was part of a previous cross-sectional clinical study of 389 Danish type 2 diabetes patients and 97 healthy controls without diabetes conducted in 2016-2018 25 ( Figure A1). These patients had participated in a large questionnaire study assessing neuropathy and neuropathic pain and were invited to participate in the clinical study. We included patients with noDPN, probable and definite DPN and all healthy controls. Subjects underwent neurological examination to establish a diagnosis of DPN. This included a detailed evaluation of symptoms and signs of neuropathy, nerve conduction studies (NCS), quantification of IENFD, quantification of corneal nerve fiber length (CNFL), fiber density (CNFD), branch density (CNBD), QST with quantification of cold detection (CDT), and warm detection (WDT) thresholds and evaluation of CAN using the Vagus™ device. The study population is described in detail previously. 25

| Modified Schellong test measuring orthostatic hypotension
Blood pressure change in response to standing was measured using a brachial blood pressure cuff to assess the presence of orthostatic hypotension. Blood pressure and heart rate were measured three times: two times after 5 min at rest while supine, and 3 min after standing. We measured blood pressure difference between the average of the two supine recordings and the recording after standing for 3 min. Orthostatic hypotension was defined as a drop in systolic blood pressure of ≥20 mmHg and/or drop in diastolic blood pressure of ≥10 mmHg of baseline within 3 min in upright position. 30,31

| COMPASS 31
The Composite Autonomic Symptom Score 31 (COMPASS 31) questionnaire was used to assess autonomic symptoms. 32 COMPASS 31 is a validated 31-item self-assessment instrument, addressing six domains of the autonomic nervous system: orthostatic, vasomotor, secretomotor, gastrointestinal, bladder, and pupillomotor functions.
The output is a global autonomic severity score and domain scores.
COMPASS 31 has previously been validated for the assessment of symptoms of autonomic neuropathy in patients with diabetes. 21 We used the validated Danish translation. 33

| NCS, IENFD, QST, and CCM
All participants had skin biopsies taken from the distal leg (10 cm above the lateral malleolus) according to international guidelines. 34 Using a bright-field immunohistochemistry protocol, intraepidermal nerve fibers were stained using PGP 9.5-antibodies where IENFD counts under the fifth centile for age and gender were considered abnormal. 34,35 NCS included examination of the sural nerve bilaterally and the median (motor and sensory), peroneal, and tibial nerves unilaterally. 36 Abnormal values were defined compared with laboratory control material according to published guidelines. [37][38][39] We assessed cold and warm detection thresholds on the dorsum of the right foot

| Ethics
All study participants gave written, informed consent and the study was approved by the Regional Research Ethics Committee of Central Denmark Region (#1-10-72-130-16). The study was registered at Aarhus University with internal notification number 62908-250. and 97 were healthy controls ( Figure A1). Characteristics of the population are described in detail elsewhere 25 as well as in the appendix (Table A1). There were no significant demographic differences between patients with DPN and noDPN. Healthy controls differed on several points from patients with diabetes; they were younger, less often male and had a lower BMI.

| Statistical analysis
Vagus™ was not performed in 44 subjects due either atrial fibrillation or pacemaker ( Figure 1) and 81 subjects were unable to perform all three tests (primarily due to insufficient expiration during Valsalva, data errors or unstable heartbeat detection). Thus, 125 subjects were excluded from Vagus™ testing and therefore from the resulting analysis of CAN. The total number of excluded subjects, and in particular the exclusion due to incomplete Vagus™, was similar in the three groups ( Figure 1). Those who were excluded were significantly older (68.6 vs. 63.5, p < 0.001) than those who were included in the analysis of CAN, but they had similar BMI (28.7 vs. 29.6, p = 0.1) and gender distribution (males excluded 50% vs. 57% included, p = 0.2).

| COMPASS 31 (autonomic symptoms)
We calculated global and domain autonomic severity scores in the neuropathy and CAN subgroups, as illustrated in Figure 3a) neuropathy groups 3b) CAN groups (only in DPN patients) and in Tables 3 and 4.
Patients with DPN reported more autonomic symptoms compared to those with noDPN (20.0 vs. 8.3, p < 0.001;  Note: Values are median (IQR) or n (%). Ever smoking (yes) was defined as either current smoking or previous smoking. Hypertension and AMI/Angina signifies current or previous disease. Both hypertension and antihypertensive medicine were reported, although due to missing data in the antihypertensive medicine question, hypertension is presented here. We can assume that most patients with diagnosed hypertension are in antihypertensive treatment, and that these variables are comparable. When deviating from the number of subjects in a group, missing data has been stated as n = x in the table. Only patients with complete Vagus™ were included in the analysis of CAN. One patient in the DPN group and 1 patient in the noDPN group were excluded due to being older than the available normal ranges for interpretation of Vagus™ tests.  Note: Values are median (IQR) or n (%). Ever smoking (yes) was defined as either current smoking or previous smoking. Hypertension and AMI/Angina signifies current or previous disease. Both hypertension and antihypertensive medicine were reported, although due to missing data in the antihypertensive medicine question, hypertension is presented here. We can assume that most patients with diagnosed hypertension are in antihypertensive treatment, and that these variables are comparable. When deviating from the number of subjects in a group, missing data has been stated as n/N or n = x in the table. Only patients with complete Vagus™ were included in the analysis of CAN. One patient in the DPN group and one patient in the noDPN group were excluded due to being older than the available normal ranges for interpretation of Vagus™ tests.

ACKNOWLEDGMENTS
We have no relevant acknowledgements.