Cardiac autonomic neuropathy (CAN) is a complication of dysglycaemia with independent prognostic significance for cardiovascular mortality1, and is now recognised to be detectable well before the diagnostic threshold for diabetes is crossed2,3. Two non-invasive methods dominate its clinical assessment: spectral analysis of heart rate variability (HRV), a passive assessment of resting autonomic tone standardised by a 1996 Task Force document4,5, and Ewing's battery of five cardiovascular reflex tests, an active assessment of the cardiovascular response to standardised physiological challenges (deep breathing, standing, Valsalva manoeuvre, and sustained handgrip)6. Although both methods are well validated individually, they interrogate the autonomic nervous system through physiologically distinct mechanisms — passive resting oscillation versus active reflex challenge — and are rarely applied concurrently in the same study population.

Early evidence from the Hoorn Study, which combined Ewing tests, HRV, and baroreflex sensitivity within a single cohort, suggested that these approaches provide complementary rather than redundant information7, and a subsequent study directly comparing simple heart-rate indices against the classical reflex battery in diabetes and prediabetes reported that combining the two approaches improved sensitivity for detecting subclinical autonomic compromise relative to either method used alone8. However, no study, to our knowledge, has directly quantified the paired, within-population difference in detection rate between a resting-HRV-based screening criterion and Ewing's battery specifically in a prediabetic population. This distinction matters practically: resting HRV recording is faster and requires less active participant cooperation than the full Ewing battery, and if the two methods were diagnostically interchangeable, HRV alone would be the more efficient screening tool. We therefore compared, within the same prediabetic cohort, the proportion of participants identified as having autonomic dysfunction by a normative HRV-based criterion against the proportion identified by Ewing's battery, using the paired nature of the data to test whether any observed

difference in detection rate was statistically significant.

2.1 STUDY POPULATION This analysis was performed within a case-control study of 100 prediabetic adults (Group I; American Diabetes Association 2024 criteria9: fasting plasma glucose 100-125 mg/dL, and/or 2-hour post-load glucose 140-199 mg/dL on a 75-g OGTT, and/or HbA1c 5.7-6.4%) and 100 age- and sex-matched healthy adults with normal glucose tolerance (Group II), recruited from a tertiary-care institution in Indore, Madhya Pradesh, India, a region in which the prediabetic pool is disproportionately large relative to overt diabetes10. Exclusion criteria were established diabetes, cardiac arrhythmia or structural heart disease, medications affecting autonomic function, non-diabetic peripheral/autonomic neuropathy, and current smoking or regular alcohol use. The study was approved by the Institutional Ethics Committee and conducted per the Declaration of Helsinki11 and ICMR National Ethical Guidelines12, with written informed consent obtained from all participants. 2.2 HRV Recording and Ewing's Battery Short-term (5-minute), supine, resting HRV was recorded using a 12-lead digital ECG/HRV-analysis system following 1996 Task Force standards4, after 10 minutes' rest, overnight fast, and 12-hour caffeine/tobacco/exercise abstention; SDNN and RMSSD were derived after artefact removal. All five components of Ewing's battery6 were performed and graded against standard cut-offs (normal/borderline/abnormal), combined into a composite score (0-5), and classified as Normal, Early, Definite, or Severe CAN. 2.3 Definition of the Comparator Criteria A normative HRV cut-off was derived directly from the study's own healthy-control (Group II) distribution, using the 5th percentile of SDNN and RMSSD, an approach methodologically consistent with normative-limit derivation used in prior large population surveys of glucose-intolerance-related autonomic dysfunction13, and arguably more locally valid than an externally imported literature threshold. A Group I participant was classified as having 'reduced HRV' if either SDNN or RMSSD fell below its respective control-derived 5th-percentile cut-off. CAN by Ewing's battery was defined as a composite score ≥1 (i.e., any grade above Normal). 2.4 Statistical Analysis The proportion of Group I participants classified as abnormal by each criterion was calculated, and the two methods were cross-tabulated at the individual-participant level to identify concordant and discordant classifications. Because both tests were applied to the same participants, the paired difference in detection rate was assessed using McNemar's exact test. Analyses were performed in a validated statistical computing environment (Python, SciPy/statsmodels); two-tailed p<0.05 was considered significant

Group I and Group II were well matched for age (median 46.0 vs 46.5 years, p=0.485) and sex distribution (67M/33F vs 56M/44F, p=0.146). The 5th-percentile healthy-control cut-offs used to define reduced HRV were SDNN 42.0 ms and RMSSD 25.0 ms.

Using this HRV-based criterion, 93 of 100 (93.0%) Group I participants were classified as having reduced HRV. Using Ewing's battery, 100 of 100 (100.0%) Group I participants were classified as having CAN of at least Early grade. Table 1 shows the paired cross-classification: 93 participants were identified as abnormal by both methods, and 7 were identified as abnormal by Ewing's battery alone; no participant was identified as abnormal by the HRV criterion but classified as normal by Ewing's battery. Overall concordance between the two methods was 93.0%. McNemar's exact test confirmed that this paired difference in detection rate was statistically significant (p=0.016; Table 1, Fig. 1).

Table 1. Paired detection of cardiac autonomic dysfunction in Group I (n=100): HRV criterion versus Ewing's battery

HRV criterion: SDNN <42.0 ms and/or RMSSD <25.0 ms (Group II 5th-percentile cut-offs). Ewing's battery: composite score ≥1. McNemar's exact test, p=0.016.

Fig. 1. Proportion of prediabetic participants (Group I, n=100) classified as having cardiac autonomic dysfunction by the HRV-based criterion versus Ewing's battery.

Among the seven discordant participants, all showed at least two abnormal or borderline heart-rate-based Ewing tests, and two showed an additional abnormal blood-pressure-based test (composite Ewing score ≥2 in all seven), while resting SDNN and RMSSD in these participants remained just above the respective control-derived cut-offs.

In this prediabetic cohort, Ewing's battery identified cardiac autonomic dysfunction in a significantly greater proportion of participants than a percentile-based short-term resting HRV criterion, with every discordant case detected by Ewing's battery alone and none by HRV alone. This finding provides direct, within-population empirical support for the combined-methods rationale first suggested by the Hoorn Study7 and subsequently demonstrated by Pafili et al. in diabetes and prediabetes8, and extends it by showing that even a rigorously derived, locally appropriate HRV cut-off, rather than a generic literature-derived threshold, can still under-detect relative to a structured, multi-domain reflex battery. A plausible physiological explanation follows from the different challenge conditions each method samples. Resting HRV predominantly captures parasympathetic (vagal) tone under a single, static physiological state4,5, whereas Ewing's battery assesses the cardiovascular reflex response to an active postural, respiratory, or isometric challenge6. A participant with borderline-preserved resting vagal tone but an inadequate reflex response to orthostatic or Valsalva challenge would be missed by resting HRV alone but correctly identified by Ewing's battery — consistent with the pattern observed in our seven discordant cases, all of whom showed multiple abnormal or borderline heart-rate-based reflex tests despite resting SDNN/RMSSD values only marginally above the control-derived cut-off. These findings have direct implications for screening design in populations with a large prediabetic burden, such as central India10. Where only one method can be deployed for reasons of time or resource constraint, our data suggest that Ewing's battery, or at minimum its heart-rate-based components (deep breathing and postural change, each requiring under two minutes), should be prioritised over resting HRV alone, or the two combined where feasible. Given the well-established prognostic significance of autonomic dysfunction detected by either method1, and the growing international evidence that such dysfunction is present from the prediabetic stage onward2,3, a normal short-term resting HRV recording should not, on the basis of these findings, be regarded as sufficient to exclude early cardiac autonomic involvement in a prediabetic individual. Limitations include the cross-sectional design; the use of a single, study-specific HRV cut-off rather than externally validated population norms, though this was a deliberate methodological choice13; the relatively small number of discordant cases (n=7), which limits further subgroup characterisation; and the use of short-term rather than 24-hour HRV recording, which may itself understate the true sensitivity of HRV-based screening relative to ambulatory monitoring

Ewing's battery detected cardiac autonomic dysfunction significantly more often than a percentile-based short-term HRV criterion in prediabetic adults, with perfect one-directional discordance favouring Ewing's battery. These findings support prioritising Ewing's battery, or combined HRV/Ewing testing where feasible, over resting HRV in isolation, when screening prediabetic populations for early cardiac autonomic neuropathy

Declarations

Funding: This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

Conflict of interest: The authors declare no conflict of interest.

Ethical approval: The study was approved by the Institutional Ethics Committee, Malwanchal University, Indore, and conducted in accordance with the Declaration of Helsinki.

Informed consent: Written informed consent was obtained from all participants.

Data availability: The datasets generated during the current study are available from the corresponding author on reasonable request.

Author contributions (CRediT): Anil Kumar: Investigation, Data curation, Formal analysis, Writing – original draft. Manila Jain: Conceptualization, Methodology, Supervision, Writing – review & editing.

1. Vinik AI, Maser RE, Mitchell BD, Freeman R. Diabetic autonomic neuropathy. Diabetes Care. 2003;26(5):1553-79.
2. Coopmans C, Zhou TL, Henry RMA, Heijman J, Schaper NC, Koster A, et al. Both prediabetes and type 2 diabetes are associated with lower heart rate variability: the Maastricht study. Diabetes Care. 2020;43(5):1126-33.
3. Eleftheriadou A, Williams S, Nevitt S, Brown E, Roylance R, Wilding JPH, et al. The prevalence of cardiac autonomic neuropathy in prediabetes: a systematic review. Diabetologia. 2021;64(2):288-303.
4. Task Force of the European Society of Cardiology and the North American Society of Pacing and Electrophysiology. Heart rate variability: standards of measurement, physiological interpretation, and clinical use. Circulation. 1996;93(5):1043-65.
5. Shaffer F, Ginsberg JP. An overview of heart rate variability metrics and norms. Front Public Health. 2017;5:258.
6. Ewing DJ, Martyn CN, Young RJ, Clarke BF. The value of cardiovascular autonomic function tests: 10 years' experience in diabetes. Diabetes Care. 1985;8(5):491-8.
7. Gerritsen J, Dekker JM, TenVoorde BJ, Bertelsmann FW, Kostense PJ, Stehouwer CD, et al. Glucose tolerance and other determinants of cardiovascular autonomic function: the Hoorn Study. Diabetologia. 2000;43(5):561-70.
8. Pafili K, Trypsianis G, Papazoglou D, Maltezos E, Papanas N. Simplified diagnosis of cardiovascular autonomic neuropathy in type 2 diabetes using Ewing's battery. Rev Diabet Stud. 2015;12(1-2):213-9.
9. American Diabetes Association Professional Practice Committee. 2. Diagnosis and classification of diabetes: Standards of Care in Diabetes-2024. Diabetes Care. 2024;47(Suppl 1):S20-42.
10. Anjana RM, Unnikrishnan R, Deepa M, Pradeepa R, Joshi SR, Bhansali A, et al. Metabolic non-communicable disease health report of India: the ICMR-INDIAB national cross-sectional study (ICMR-INDIAB-17). Lancet Diabetes Endocrinol. 2023;11(7):474-89.
11. World Medical Association. World Medical Association Declaration of Helsinki: ethical principles for medical research involving human subjects. JAMA. 2013;310(20):2191-4.
12. Indian Council of Medical Research. National Ethical Guidelines for Biomedical and Health Research Involving Human Participants. New Delhi: ICMR; 2017.
13. Ziegler D, Voss A, Rathmann W, Strom A, Perz S, Roden M, et al; KORA Study Group. Increased prevalence of cardiac autonomic dysfunction at different degrees of glucose intolerance in the general population: the KORA S4 survey. Diabetologia. 2015;58(5):1118-28.