Long-term prognostic value of heart-rate recovery after treadmill testing in patients with diabetes mellitus

International Journal of Cardiology 134 (2009) 67 – 74 www.elsevier.com/locate/ijcard

Long-term prognostic value of heart-rate recovery after treadmill testing in patients with diabetes mellitus Panagiotis Georgoulias a,⁎, Nikolaos Demakopoulos b , Varvara Valotassiou a , Alexandros Orfanakis c , Alexia Zaganides a , Ioannis Tsougos d , Ioannis Fezoulidis a a

Department of Nuclear Medicine, Medical School, University of Thessaly and University Hospital of Larissa, Larissa, Greece b Department of Nuclear Medicine, NIMTS Hospital, Athens, Greece c Department of Cardiology, University Hospital of Heraklion, Heraklion, Greece d Department of Medical Physics, Medical School, University of Thessaly, Larissa, Greece Received 24 June 2007; received in revised form 17 October 2007; accepted 14 January 2008 Available online 21 May 2008

Abstract Background: Heart-rate recovery (HRR) is considered to be an independent predictor of cardiac and all-cause mortality. We examined the long-term prognostic value of HRR in patients suffering from diabetes mellitus. Methods: In this study, we included 258 consecutive patients. Patients whose HRR value or myocardial perfusion imaging could have been influenced by factors other than myocardial ischaemia, were excluded. The value of HRR was defined as the decrease in the heart-rate from peak exercise to 1 min after the termination of the exercise. All patients underwent SPECT myocardial perfusion imaging combined with exercise testing. Cardiovascular death and non-fatal myocardial infarction were considered as hard cardiac events, while late revascularization procedures as soft events. Cox proportional-hazard models were applied to evaluate the association between HRR and the investigated outcome. Results: During the follow-up period (30.8 ± 6.9 months), hard cardiac events occurred in 21 (8%) patients (15 with abnormal HRR value, p b 0.001), while 35 (14%) patients underwent revascularization (31 with abnormal HRR value, p b 0.001). Considering it as a continuous variable, HRR was a strong predictor for both hard cardiac (coefficient = −0.41, SE = 0.052, p b 0.001) and soft cardiac events (coefficient = −0.63, SE = 0.058, p b 0.001). After adjustments were made for potential confounders, including scintigraphic variables, abnormal HRR remained an independent predictor for hard and soft cardiac events (p b 0.001). Conclusion: Our results suggest that among patients with diabetes, a decreased HRR is a significant independent predictor of hard and soft cardiac events. © 2008 Elsevier Ireland Ltd. All rights reserved. Keywords: Heart-rate recovery; Diabetes; Treadmill testing; Myocardial SPECT; Prognosis

1. Introduction The rise in heart-rate during exercise is considered to be due to a combination of parasympathetic withdrawal and ⁎ Corresponding author. 15 Kritis St. 19009 Rafina, Greece. Tel.: +30 22940 26201, +30 2410 682052; fax: +30 22940 78400, +30 2410 670117. E-mail addresses: [email protected], [email protected] (P. Georgoulias). 0167-5273/$ - see front matter © 2008 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.ijcard.2008.01.036

sympathetic activation, whereas the decline in heart-rate immediately after exercise has been proposed to be a function of the reactivation of the parasympathetic tone [1–3]. The autonomic nervous system is responsible for significant influences on myocardial pathophysiology, while autonomic neuropathy is a common complication in diabetic patients and particularly in those using insulin [4,5]. The pathophysiology of abnormal heart-rate recovery (HRR) involves the inability to slow down the heart-rate

68

P. Georgoulias et al. / International Journal of Cardiology 134 (2009) 67–74

immediately after exercise, which has been previously shown to be a marker of decreased vagal activity. Attenuated HRR after exercise testing has been found to be an independent predictor of cardiac mortality in patients referred for exercise electrocardiogram [6–11]. Although diabetes mellitus is associated with increased mortality from cardiovascular disease, only few published data have engaged with the predictive value of a decreased HRR in diabetic patients [11–13]. The aim of this study was to evaluate the usefulness of HRR after treadmill testing as a long-term prognostic marker of cardiovascular morbidity and mortality in diabetic patients, in comparison to the prognostic value of exercise testing and myocardial single photon emission computed tomography (SPECT) variables. 2. Materials and methods 2.1. Population The study cohort consisted of 258 consecutive diabetic patients (168 men and 90 women) ranging in age from 33 to 81 years (mean age 55.69 ± 11.24 years), who were properly referred (based on their data), between May 2002 and December 2004, for a symptom-limited exercise testing combined with a SPECT myocardial perfusion imaging, for the evaluation of known or suspected coronary artery disease and patient risk stratification (Table 1). All patients were suffering from diabetes mellitus according to the American Diabetes Association criteria [14]. Only patients using hypoglycemic medication were included in the study; 109 patients were using insulin and 149 were using oral hypoglycemic agents (Table 1). We excluded pregnant women, patients with a history of heart failure, left bundle branch block, pre-excitation syndromes, atrioventricular block or known sick sinus syndrome, with atrial fibrillation or other tachyarrhythmias, with Table 1 Characteristics of the study group in comparison to heart-rate recovery value. Characteristic

Number of patients Age (years) Sex (male) — no. (%) Smoking — no. (%) Obesity — no. (%) Hypertension — no. (%) Lipid disorder — no. (%) Glucose (mg/dl) Use of insulin — no. (%) Use of β-blockers — no. (%) Use of calcium – channel antagonists – no. (%) Use of nitrates — no. (%) Use of other antiarrhythmics — no. (%)

Normal HRR

Abnormal HRR

(≥21 bpm)

(b21 bpm)

179 51 ± 13 108 (60%) 77 (43%) 62 (35%) 87 (49%) 103 (58%) 118 ± 6 61 (34%) 33 (18%) 37 (21%)

79 63 ± 11⁎⁎⁎ 60 (76%)⁎⁎ 53 (67%)⁎⁎⁎ 41 (52%)⁎⁎ 48 (61%)⁎ 69 (87%)⁎⁎⁎ 135 ± 9⁎⁎⁎ 42 (53%)⁎⁎ 31 (39%)⁎⁎ 35 (44%)⁎⁎⁎

29 (16%) 3 (0.016%)

34 (43%)⁎⁎⁎ 2 (0.025%)

HRR, heart-rate recovery; bpm, beats per minute. ⁎p b 0.05, ⁎⁎p b 0.01, ⁎⁎⁎p b 0.001.

bradycardia, those with a history of prior myocardial infarction or a completely irreversible (permanent) defect in their scintigram, previous cardiac surgery (bypass grafting or angioplasty), congenital or valvular heart disease, patients with cardiomyopathy and those with an implanted pacemaker. We also excluded patients taking digoxin or amiodarone (due to their prolonged chronotropic effect) and those with contraindication to or inability to perform treadmill testing or to achieve a satisfactory exercise level because of an exocardiac condition (peripheral vascular disease, sciatica, neuropathy, disability etc.). Medications that could possibly influence patient performance on exercise testing and the related variables, were temporarily withdrawn (for about five half-lives). β-blockers were discontinued gradually (within a week — depended on the medication and the dose), with complete discontinuation at least 48 h before and during the study. Calcium channel antagonists and nitrates were discontinued 48 and 24 h before and during the study, respectively. Other antiarrhythmic medications were also discontinued (at least 48 h before and during the study, according to medication half half-life). Additionally, any patient whose medication had not been discontinued as described above was excluded. Finally, patients who were lost to follow-up or died of a non-cardiovascular cause during follow-up and those who underwent early revascularization (b3 months after myocardial SPECT: 12 CABG and 10 PCI), were excluded from the prognostic analysis. Before testing, all patients gave informed consent for their participation, according to the Hospital Ethics Committee guidelines (based on the ethical guidelines of the 1975 Declaration of Helsinki), and a brief structured interview during which we obtained data on symptoms, medications, previous cardiac events, coronary risk factors and cardiac or non-cardiac diagnoses. All patients had fasting blood glucose measurements taken prior to the study. Hypertension was considered as a systolic blood pressure of 140 mm Hg or greater at rest and/ or a diastolic blood pressure of 90 mm Hg or greater at rest, or treatment with antihypertensive medicines. Diagnoses of lipid disorders were derived from the interviews with the patients and the use of relevant medications. Obesity was considered as a body mass index (BMI — calculated as weight in kilograms divided by height in meters squared) of 30.0 or greater. Prior to the study, patients were also given written directions on radioprotection. 2.2. Follow-up Follow-up data were obtained by phone contact with the patients, their relatives and patients' general practitioner or cardiologist, during visits to the clinic and/or review of the patients' hospital records. Cardiovascular death and nonfatal myocardial infarction were considered as hard cardiac events, while revascularization (≥ 3 months after myocardial

P. Georgoulias et al. / International Journal of Cardiology 134 (2009) 67–74

SPECT, including PCI and CABG) as soft events. Death was considered cardiovascular if it was caused by “Diseases of the Circulatory System” according to the Tenth International Classification of Diseases (numbers I00–I99) [15]. In the presence of multiple causes, a hierarchical preference was adopted with cancer in advanced stages and cardiovascular death in that order [16]. A reviewer, who was blinded to clinical, exercise, myocardial SPECT data and the hypothesis of the study, assessed the cause of death by examination of death certificates. All patients included in the study were followed up for at least 7 months. The follow-up period was completed in patients suffering from a hard or soft event. The average follow-up period was 30.8 ± 6.9 months (range 7–44 months). 2.3. Exercise testing After the discontinuation of cardioactive medication, a 12 hour fasting, avoiding smoking or engaging in heavy physical activity for at least 3 h before the examination and modification of antidiabetic medication to avoid hypoglycemic symptoms, the patients underwent a symptom-limited treadmill exercise testing (Bruce protocol), as described previously [17,18]. Data on symptoms and estimated workload in metabolic equivalents (METs) were obtained. Functional capacity in METs was estimated using standard tables [19]. As a criterion of ischaemic ST-segment response, we considered greater than 1 mm horizontal or downsloping STsegment depression 80 ms after the J point or more than 1 1 mm of additional ST-segment rise in leads without pathologic Q waves. Chronotropic response, heart-rate reserve (%) and Duke's treadmill score were calculated according to standard formulas [17,18,20–22]. After achieving peak workload, the treadmill stopped (no “cool-down” period was allowed) and the patients immediately laid on a bed, placed next to the treadmill. The patients remained in supine position for a period of 8 min, which was considered the recovery period (this period was prolonged in case the symptoms or electrocardiographic changes were persistent). The reduction in heart-rate from its value at peak exercise to the rate 1 1 min later was determined as the HRR [6–9,17,18]. 2.4. SPECT myocardial perfusion imaging Myocardial perfusion exercise-rest studies were performed using thallium-201 (Tl-201) or technetium-99m (Tc-99m) tetrofosmin (Myoview, Amersham). One hundred and fifty-four patients (60%) were studied with Tl-201. One hundred and four patients (40%), especially obese patients or women with large breasts, were studied with Tc-99m tetrofosmin (1-day protocol, high-count rest scans were acquired as gated-SPECT), in order to minimize the effect of soft-tissue attenuation artifacts. Acquisition and processing protocols (Sopha Medical Vision software — Quantitative Gated SPECT) used in our Department for Tl-201 and Tc-

69

99m-tetrofosmin SPECT studies have been described in detail elsewhere, based on EANM/ESC procedural guidelines [17,18,23–25]. We performed polar and 3-dimensional mapping in all studies (with Tl-201 or Tc-99m tetrofosmin) and calculated the transient ischaemic dilation (TID) index. For SPECT interpretation, the myocardium of the left ventricle was divided into 17 segments, according to previous reports [25,26]. Two independent experienced observers blindly evaluated the reconstructed images, the polar maps and the 3-dimensional images of both stress and rest studies by scoring uptake in each of the 17 regions, using a 5-point scoring system (0: normal uptake, 1: mildly reduced uptake, 2: moderately reduced uptake, 3: severely reduced uptake and 4: no uptake) [25,26]. If counts were reduced in a segment and this was judged to be the result of attenuation artefact, the score was zero [25]. Ischaemia was considered in every region of the myocardium with an uptake higher than zero at stress imaging and a reduction of the score by at least one unit at rest. Finally, the “summed stress score” (SSS) and “summed rest score” (SRS) were obtained by adding the scores of the regions in stress and rest studies [25–27]. In 16 (6.2%) studies (11 with Tl-201 and 5 with Tc-99m-tetrofosmin), where discordance between the two observers was detected (difference N 2 in SSS and/or SRS values), the view of a third observer was requested and the disagreement was resolved by consensus. Subsequently, a “summed difference score” (SDS) was estimated by subtracting the SRS from the SSS to assess defect reversibility [25–27]. Studies with an SSS equal or lower than 2 were considered normal. 2.5. Statistical analysis Continuous variables are presented as mean ± standard deviation (SD). In univariate analysis the potential confounders included age, gender, smoking, fasting blood glucose level, the presence or absence of chest pain, hypertension, obesity, lipid disorder; the use or non-use of cardioactive medications (β-blockers, calcium channel antagonists, nitrates) and insulin, resting heart-rate, exercise duration, maximal systolic blood pressure, maximal heart-rate, METs, double product, Duke's score, angina and abnormal ST response during exercise testing. Cox proportional-hazards regression analysis was applied to determine the independent predictors for hard and soft cardiac events. The prognostic value of each of the aforementioned variables was first assessed by univariate Cox regression analysis. Variables that showed significant association with the outcome were included in the multivariate Cox proportional-hazard model in a forward stepwise method. The proportional proportional-hazards assumption was verified by inspection of log(− log[survival function]) curves. Unadjusted cardiac event-free survival estimates were calculated using the Kaplan–Meier method, while adjusted

70

P. Georgoulias et al. / International Journal of Cardiology 134 (2009) 67–74

Table 2 Exercise performance and myocardial SPECT data. Data

Resting heart-rate (bpm) Exercise duration (min) Maximal heart-rate (bpm) HRR/maximal heart-rate (%) Heart-rate reserve (%) Total time to recovery (min) Maximal systolic blood pressure (mm Hg) Metabolic equivalents (METs) Double product (×100) Duke's score Angina during exercise testing — no. (%) Abnormal ST-segment response — no. (%) Abnormal chronotropic response — no. (%) Myocardial SPECT with Tl-201 — no. (%) Myocardial SPECT with Tc-99m-tetrofosmin — no. (%) Myocardial perfusion SSS Myocardial perfusion SRS Myocardial perfusion SDS TID index

Normal HRR

Abnormal HRR

(≥21 bpm)

(b21 bpm)

68 ± 10 11.2 ± 3.1 154 ± 13 23.1 ± 3.2 87 ± 15 5.5 ± 0.46 203 ± 14

91 ± 11⁎⁎⁎ 8.1 ± 2.2⁎⁎⁎ 112 ± 18⁎⁎⁎ 16.8 ± 2.4⁎⁎⁎ 61 ± 12⁎⁎⁎ 6.7 ± 0.6⁎⁎⁎ 164 ± 17⁎⁎⁎

11.7 ± 3.1 306 ± 34 5.6 ± 2.4 17 (9%)

8.6 ± 2.7⁎⁎⁎ 231 ± 38⁎⁎⁎ − 2.3 ± 1.7⁎⁎⁎ 22 (28%)⁎⁎

42 (23%)

34 (43%)⁎⁎

56 (31%)

45 (57%)⁎⁎

105 (59%)

49 (62%)

74 (41%)

30 (38%)

6.4 ± 2.3 2.0 ± 0.93 6.1 ± 1.4 0.96 ± 0.14

18.9 ± 4.8⁎⁎⁎ 2.3 ± 1.6 16.2 ± 3.9⁎⁎⁎ 1.08 ± 0.21⁎⁎⁎

HRR, heart-rate recovery; bpm, beats per minute; SSS, summed stress score; SRS, summed rest score; SDS, summed difference score; TID, transient ischaemic dilation. ⁎p b 0.05, ⁎⁎p b 0.01, ⁎⁎⁎p b 0.001.

survival functions were calculated and plotted according to a likelihood-based approach estimation of the baseline survival function. Cardiac event-free survival curves were compared using the log-rank test. All analyses were performed using Minitab 13 (http:// www.minitab.com). In all analyses, a p-value b 0.05 was considered statistically significant. 3. Results A total of 258 patients participated in the study. The mean value of HRR was approximately 31.2 ± 12.6 beats/min with a range from 8 to 67 beats/min.

Seventy-nine patients (31%) had an abnormal HRR value (b 21 beats/min, according to a previously conducted study) during the first minute after exercise testing cessation [17]. In addition, 211 patients (82%) had an abnormal myocardial perfusion SPECT. Only 5 patients (6%) out of 79 with an abnormal HRR value had a normal myocardial perfusion study. The main characteristics of the patients according to the value of their HRR (normal or abnormal value) are presented in Table 1. In general, patients with abnormal HRR value, had a higher frequency of other risk factors (besides diabetes) for CAD, had significantly higher mean fasting blood glucose level and were mostly taking cardioactive medications and using insulin. The heart-rate percent change from peak exercise (1 min after peak) was 23.1 ± 3.2 for patients with normal HRR value, compared to 16.8 ± 2.4 for patients with abnormal HRR value (p b 0.001, Table 2). Patients with an abnormal HRR value had lower efficiency during treadmill testing, and presented more pathologic findings on the scintigram, compared to subjects with a normal HRR value (Table 2). During the follow-up period (30.8 ± 6.9 months), hard cardiac events occurred in 21 (8%) patients (15 with abnormal HRR value, p b 0.001). Specifically, cardiac death occurred in 6 patients (5 with abnormal HRR value, p b 0.001) and non-fatal myocardial infarction in 15 patients (12 with abnormal HRR value, p b 0.001). In addition, 35 (14%) patients underwent revascularization (31 with abnormal HRR value, p b 0.001). The time interval between patient examination and cardiac events was 23.5 ± 5.7 months for hard and 27.3 ± 8.2 months for soft events. Univariate Cox proportional-hazard regression analyses showed that HRR, when considered as a continuous variable, was a strong predictor for both hard cardiac (coefficient = − 0.41, SE = 0.053, p b 0.001) and soft cardiac events (coefficient = − 0.63, SE = 0.058, p b 0.001). The HRR, gender, exercise duration, angina during exercise testing, Duke's score, SSS, SDS and the use of nitrates and insulin, showed a significant association with the outcome and were included in the multivariate Cox proportional-hazard regression analysis for hard cardiac events. Angina during exercise testing, HRR, SSS, SDS and the use of nitrates and insulin were included in the multivariate analysis for soft cardiac events. After fitting the multivariate Cox regression analysis, HRR, the use of nitrates, insulin, SDS and SSS were independent

Table 3 Multivariate Cox regression analysis for prediction of hard and soft cardiac events. Hard cardiac events

Nitrates Insulin SDS SSS HRR

Soft cardiac events

Coeff.‡

SE‡

95% CI†

Wald test

p

Coeff.

SE

95% CI

Wald test

p

1.264 1.158 0.136 0.461 −0.295

0.487 0.522 0.029 0.172 0.052

0.31–2.21 0.14–2.18 0.08–0.19 0.80–0.12 − 0.39–(− 0.19)

6.74 4.92 21.55 7.18 31.10

0.009 0.026 b0.001 b0.001 b0.001

3.0811 – 0.154 – − 0.233

1.1115 – 0.0207 – 0.058

0.90–5.26

7.68

0.11–0.19

55.35

− 0.35–(− 0.12)

16.14

0.005 – b0.001 – b0.001

‡Coefficient; ‡Standard error; †95% Confidence Interval. HRR, heart-rate recovery; SSS, summed stress score; SDS, summed difference score.

P. Georgoulias et al. / International Journal of Cardiology 134 (2009) 67–74

71

Fig. 1. Kaplan–Meier curves of event-free survival according to abnormal heart-rate recovery for soft cardiac events. Dotted line is for patients with HRR ≥21; continuous line is for patients with HRR b21 (log rank: 117.8, p b 0.001).

predictors of hard cardiac events, whereas only an abnormal value of HRR, the use of nitrates and SDS, were independent predictors of soft cardiac events (Table 3). Patients were divided into two groups according to their HRR score (i.e. ≥ 21 and b21). Univariate Cox proportionalhazard regression analyses showed that HRR, when coded as mentioned above, was a strong predictor for both hard cardiac (coefficient = − 3.18, SE = 0.626, p b 0.001) and soft cardiac events (coefficient = − 3.6, SE = 0.533, p b 0.001). When the two groups were compared to each other using cardiac event-free Kaplan–Meier curves of survival, for both soft and hard cardiac events and the log-rank test, we confirmed that patients with an abnormally low HRR value had significantly decreased event-free survival (p b 0.001, Figs. 1 and 2). When a separate multiple analysis for subjects under insulin therapy was performed, it was found that the independent predictors for soft events were SDS (coefficient =

0.104, SE = 0.027, p b 0.001) and HRR (coefficient = − 0.226, SE = 0.063, p = 0.005). Additionally, for hard events the corresponding independent predictors were: SSS (coefficient = 0.522, SE = 0.224, p = 0.020), SDS (coefficient = 0.505, SE = 0.245, p = 0.040) and HRR (coefficient = − 0.267, SE = 0.088, p = 0.003). 4. Discussion This study further supports the prognostic value of HRR and demonstrates that diabetic patients with a decreased HRR have increased likelihood of cardiovascular death, nonfatal myocardial infraction and revascularization procedures (CABG, PCI). The prognostic importance of the HRR after exercise has already been well established in general populations [6– 10,28–30]. An attenuated HRR after treadmill testing seems to be an important predictor of overall mortality, independent

Fig. 2. Kaplan–Meier curves of event-free survival according to abnormal heart-rate recovery for hard cardiac events. Dotted line is for patients with HRR ≥21; continuous line is for patients with HRR b21 (log rank: 55.3, p b 0.001).

72

P. Georgoulias et al. / International Journal of Cardiology 134 (2009) 67–74

of the workload, the presence or absence of myocardial perfusion defects and the changes in heart-rate during exercise [6–10,28–30]. Cardiovascular disease is the major cause of death in patients suffering from diabetes mellitus and is far more prevalent than in the non-diabetic population because of accelerated atherogenesis, while heart-rate variability is reduced in the presence of autonomic dysfunction that is related with diabetes mellitus [4,12,13,31–33]. In addition, other investigators have reported that fasting plasma glucose and plasma insulin levels are closely correlated with changes in cardiac autonomic function and impaired HRR [5,11,28,34–36]. It seems that elevated fasting plasma levels of glucose are associated with a progressive decrease in vagal tone which is primarily responsible for the decline in heart-rate post-exercise [5,28,37]. Previous studies have postulated that persistently elevated levels of glucose, damage peripheral nerve fibers thus stimulating sympathetic activity and decreasing parasympathetic control [34,35,37]. Furthermore some groups have suggested that insulin levels could be correlated with autonomic neuropathy, potentially affecting cardiac electric activity [29,36]. Another possible explanation for the association between the use of insulin and attenuated HRR, could be that patients using insulin commonly have longstanding elevated glucose levels, usually before starting insulin therapy. Moreover, the association between poor autonomic function and diabetes has already been reported [5,28,31–33]. In the presence of autonomic dysfunction the parasympathetic activity is altered, manifested by a decrease in vagal tone and hence an increase in heart-rate at rest, which can be considered as a global marker of cardiac autonomic function [5,28,39]. This finding has also been reported in our study in patients with an abnormal value of HRR, compared to patients with a normal value (Table 2). On the other hand, there are only few data on the prognostic importance of HRR in diabetic patients, although Cheng et al. [11] in a cohort study of men with diabetes, suggested that patients with impaired HRR (defined as the heart-rate decline during the first 5 5 min after the completion of the maximal exercise test) had a higher risk of cardiovascular and all-cause mortality. In a previously conducted study, using the same method to perform exercise testing and to calculate the HRR value in a healthy cohort of subjects, we defined a heart-rate decrease of 21 beats/min, 1 min after the cessation of peak exercise, as the lowest normal value [17]. Although in other studies an HRR value of more than 12 beats/min has been proposed to be normal, the conflicting results could be attributed to differences in methodology [6,7,10]. Patients in our study, did not undergo a “cool-down” period, but they terminated the testing and laid down almost immediately after achieving the peak of the exercise, in accordance with other studies [9,39,40]. We believe that HRR may be affected, even by the small workload during the “cool-down” period. Moreover, the implication of the workload during this period is different

for each patient, since it is effort-dependent. Most probably, this is an important reason why the lowest normal value for HRR defined using the aforementioned methodology was significantly greater than the values reported in other studies with a cool-down period post-treadmill exercise [6,7,9,10,17,39,40]. The main point is that patients in our study, comparable to Watanabe et al. did not undergo a “cool-down” period after achieving the peak of the exercise, as we have described in detail previously [6,7,9,10,17,18]. We have recently reported the relationship between the HRR and myocardial ischaemia in diabetic patients, as it is mainly assessed via myocardial perfusion SPECT variables (SSS, SDS, p b 0.001) and also between HRR and the chronotropic variables (peak heart-rate, chronotropic response, heart-rate reserve, p b 0.001) [18]. This reported association is in general accordance with other reports despite the fact that the majority of them did not exclusively include patients with diabetes. Nevertheless, conflicting results have also been published regarding the correlation between HRR and myocardial perfusion parameters [6–11,38,39]. The mechanism by which myocardial ischaemia could affect HRR is not clear. However, it has been previously reported that the effect of myocardial ischaemia on heart neurosis, results in nerve injury which may lead to electrical imbalance [41]. In our study, a low value of HRR was independently predictive of both hard and soft cardiac events (p b 0.001). The association between HRR and hard or soft cardiac events persisted even after adjusting for common cardiovascular risk factors and accounting for other possible confounders. These data are in concordance with the results of Cheng et al. [11], despite the fact that they included only diabetic men in their study. Furthermore, the prognostic value of HRR was independent of the exercise testing parameters and also of the myocardial SPECT data (SSS, SRS, SDS, TID), findings which indicate that the prognostic value of HRR can not be attributed only to its correlation with myocardial ischaemia and the related parameters. To our knowledge, this is the first report studying the predictive value of HRR in diabetics, adjusting for well established myocardial SPECT prognostic factors. The mechanism by which impaired HRR confers an increased risk of cardiac events could be explained by the fact that the fall in heart-rate immediately after exercise has been considered to be a function of parasympathetic nervous system reactivation, thus an attenuated HRR is indicative of autonomic dysfunction [6,28,33]. It is also well known that increased vagal activity has been associated with decreased risk of death, while dysfunctional autonomic heart-rate response has been related to lethal cardiac arrhythmias, a hypotheses that could suggest a pathophysiologic explanation at least for the observed independent association between HRR and hard cardiac events [6,28,33].

P. Georgoulias et al. / International Journal of Cardiology 134 (2009) 67–74

5. Potential limitations Our study has some potential limitations. First, the percentage of patients with an abnormal HRR value was relatively low (33%), although it was similar to that of other studies [5–7,11,28]. Second, we did not examine all patients using the same tracer, so small differences in defect scores related to isotope properties can not be entirely excluded, although the semiquantitative analysis we used is considered reliable for both Tl-201 and Tc-99m labeled tracers [24,26]. Moreover, the independent prognostic value of lung/heart ratio (LHR) and LVEF was not assessed, since these parameters were calculated only in patients examined with Tl-201 and Tc99m tetrofosmin respectively. Third, we did not have measures of maximum VO2 so we did not include a direct estimation of patient physical fitness in our analysis. However, METs calculation is considered to be a reliable alternative in estimating cardiorespiratory fitness [11,19]. Finally, in contrast to the study of Cheng et al., we did not include deaths of a non-cardiovascular origin in hard events, to avoid blurred results. However, we consider that our results are biased to null as the assessment of the cause of death was blinded to clinical, exercise, myocardial SPECT data and the hypothesis of the study. 6. Conclusions — clinical implications In the present study we evaluated the long-term prognostic value of HRR in patients with diabetes mellitus. In conclusion, our results suggest that HRR is an independent and powerful predictor for hard cardiac events (death or myocardial infarction) and also for revascularization procedures (soft events), in diabetic patients. It seems that the calculation of HRR value, which is a simply-calculated marker during exercise testing, will maximize not only the information it provides for the assessment of the severity of myocardial ischaemia, but also the prognostic value of the study, especially in diabetic patients who are considered as high-risk for cardiovascular disease. Acknowledgment The authors thank the MSc Theofilos Topaltzikis and the technologists Stavroula Giannakou, Anastasia Ziaka and Socrates Kapikos for their contribution to the present study. References [1] Ellestad MH. Chronotropic incompetence. The implications of heart rate response to exercise (compensatory parasympathetic hyperactivity?). Circulation 1996;93:1485–7. [2] Arai Y, Saul JP, Albrecht P, et al. Modulation of cardiac autonomic activity during and immediately after exercise. Am J Physiol 1989;256: H 132–41.

73

[3] Imai K, Sato H, Hori M, et al. Vagally mediated heart rate recovery after exercise is accelerated in athletes but blunted in patients with chronic heart failure. J Am Coll Cardiol 1994;24: 1529–35. [4] May O, Arildsen H, Damsgaard EM, Mickley H. Cardiovascular autonomic neuropathy in insulin-dependent diabetes mellitus: prevalence and estimated risk of coronary heart disease in the general population. J Intern Med 2000;248:483–91. [5] Seshadri N, Acharya N, Lauer MS. Association of diabetes mellitus with abnormal heart rate recovery in patients without known coronary artery disease. Am J Cardiol 2003;91:108–11. [6] Cole CR, Blackstone EH, Pashkow FJ, Snader CE, Lauer MS. Heartrate recovery immediately after exercise as a predictor of mortality. N Engl J Med 1999;341:1351–7. [7] Desai MY, De La Pena-Almaguer E, Mannting F. Abnormal heart rate recovery after exercise: a comparison with known indicators of increased mortality. Cardiology 2001;96:38–44. [8] Cole CR, Foody JM, Blackstone EH, Lauer MS. Heart rate recovery after submaximal exercise testing as a predictor of mortality in a cardiovascularly healthy cohort. Ann Intern Med 2000;132:552–5. [9] Watanabe J, Thamilarasan M, Blackstone EH, Thomas JD, Lauer MS. Heart rate recovery immediately after treadmill exercise and left ventricular systolic dysfunction as predictors of mortality; the case of stress echocardiography. Circulation 2001;104:1911–6. [10] Nishime EO, Cole CR, Blackstone EH, Pashkow FJ, Lauer MS. Heart rate recovery and treadmill exercise score as predictors of mortality in patients referred for exercise ECG. JAMA 2000;284:1392–8. [11] Cheng YJ, Lauer MS, Earnest CP, et al. Heart rate recovery following maximal exercise testing as a predictor of cardiovascular disease and all cause mortality in men with diabetes. Diabetes Care 2003;26: 2052–7. [12] Nesto RW. Correlation between cardiovascular disease and diabetes mellitus: current concepts. Am J Med 2004;116:11S–22S. [13] Coutinho M, Gerstein HC, Wang Y, Yusuf S. The relationship between glucose and incident cardiovascular events. A meta regression analysis of published data from 20 studies of 95,783 individuals followed for 12.4 years. Diabetes Care 1999;22:233–40. [14] American Diabetes Association. Diagnosis and classification of diabetes mellitus. Diabetes Care 2004;27:S5–S10. [15] Coster CD, Quan H, Finlayson A, et al. Identifying priorities in methodological research using ICD-9-CM and ICD-10 administrative data: report from an international consortium. BMC Health Services Res 2006;6:77. [16] Pitsavos CH, Chrysohoou C, Panagiotakos DB, et al. Exercise capacity and heart rate recovery as predictors of coronary heart disease events, in patients with heterozygous familial hypercholesterolemia. Atherosclerosis 2004;173:347–52. [17] Georgoulias P, Orfanakis A, Demakopoulos N, et al. Abnormal heart rate recovery immediately after treadmill testing: correlation with clinical, exercise testing, and myocardial perfusion parameters. J Nucl Cardiol 2003;10:498–505. [18] Georgoulias P, Demakopoulos N, Orfanakis A, et al. Evaluation of abnormal heart-rate recovery after exercise testing in patients with diabetes mellitus: correlation with myocardial SPECT and chronotropic parameters. Nucl Med Commun 2007;28:165–71. [19] Fletcher GF, Balady G, Froelicher VF, Hartley LH, Haskell WL, Pollock ML. Exercise standards. A statement for healthcare professionals from the American Heart Association Writing Group. Circulation 1995;91:580–615. [20] Lauer MS, Francis GS, Okin PM, Pashkow FJ, Snader CE, Marwick TH. Impaired chronotropic response to exercise stress testing as a predictor of mortality. JAMA 1999;281:524–9. [21] Rahimi K, Thomas A, Adam M, Hayerizadeh BF, Schuler G, Secknus MA. Implications of exercise test modality on modern prognostic markers in patients with known or suspected coronary artery disease: treadmill versus bicycle. Eur J Cardiovasc Prev Rehabil 2006;13: 45–50.

74

P. Georgoulias et al. / International Journal of Cardiology 134 (2009) 67–74

[22] Mark DB, Shaw L, Harrell Jr FE, et al. Prognostic value of a treadmill exercise score in outpatients with suspected coronary artery disease. N Engl J Med 1991;325:849–53. [23] Georgoulias P, Demakopoulos N, Kontos A, et al. 99mTc-tetrofosmin myocardial perfusion imaging: a comparison with coronary angiography. Nuklearmedizin 1996;35:153–5. [24] Hesse B, Tägil K, Cuocolo A, et al. EANM/ESC procedural guidelines for myocardial perfusion imaging in nuclear cardiology. Eur J Nucl Med 2005;32:855–97. [25] Kapur A, Latus KA, Davies G, et al. A comparison of three radionuclide myocardial perfusion tracers in clinical practice: the ROBUST study. Eur J Nucl Med 2002;29:1608–16. [26] Cerquira MD, Weissman NJ, Dilsizian V, et al. Standardized myocardial segmentation and nomenclature for tomographic imaging of the heart: a statement for healthcare professionals from the Cardiac Imaging Committee of the Council on Clinical Cardiology of the American Heart Association. J Nucl Med 2002;9:240–5. [27] Schinkel AFL, Elhendy A, Van Domburg RT, Bax JJ, Vourvouni EC, Bountioukos M. Incremental value of exercise technetium-99m tetrofosmin myocardial perfusion single-photon emission computed tomography for the prediction of cardiac events. Am J Cardiol 2003;91:408–11. [28] Panzer C, Lauer MS, Brieke A, Blackstone E, Hoogwerf B. Association of fasting plasma glucose with heart rate recovery in healthy adults. A population-based study. Diabetes 2002;51:803–7. [29] Mora S, Redberg RF, Cui Y, et al. Ability of exercise testing to predict cardiovascular and all-cause death in asymptomatic women: a 20-year follow-up of the lipid research clinics prevalence study. JAMA 2003;290:1600–7. [30] Morise AP. Heart rate recovery: predictor of risk today and target of therapy tomorrow? Circulation 2004;110:2778–80. [31] Malik M, Camm AJ. Components of heart rate variability — what they really mean and what we really measure. Am J Cardiol 1993;72:821–2.

[32] Tsuji H, Venditti FJ, Manders ES, et al. Determinants of heart rate variability. J Am Coll Cardiol 1996;28:1539–46. [33] Singh JP, Larson MG, O CJ, Donnell, et al. Association of hyperglycemia with reduced heart rate variability (The Framingham Heart Study). Am J Cardiol 2000;86:309–12. [34] Aso Y, Fujiwara Y, Inukai T, Takemura Y. Power spectral analysis of heart rate variation in diabetic patients with neuropathic foot ulceration. Diabetes Care 1998;21:1173–7. [35] Emdin M, Gastaldelli A, Muscelli E, et al. Hyperinsulinemia and autonomic nervous system dysfunction in obesity: effects of weight loss. Circulation 2001;103:513–9. [36] Van De Borne P, Hausberg M, Hoffman RP, Mark AL, Anderson EA. Hyperinsulinemia produces cardiac vagal withdrawal and nonuniform sympathetic activation in normal subjects. Am J Physiol 1999;276: R 178– 183. [37] Carnethon MR, Jacobs DR, Sidney S, Liu K. Influence of autonomic nervous system dysfunction on the development of type 2 diabetes. The Cardia study. Diabetes Care 2003;26:3035–41. [38] Lima RS, Lorenzo AD, Soares AJ. Relation between postexercise abnormal heart rate recovery and myocardial damage evidenced by gated single-photon emission computed tomography. AJC 2006;97: 1452–4. [39] Nonaka A, Shiotani H, Kitano K, Yokoyama M. Determinants of heart rate recovery in patients with suspected coronary artery disease. Kobe J Med Sci 2007;53:93–8. [40] Bilsel T, Terzi S, Akbulut T, Sayar N, Hobikoglu G, Yesilcimen K. Abnormal heart rate recovery immediately after cardiopulmonary exercise testing in heart failure patients. Int Heart J 2006;47:431–40. [41] Lautamäki R, Dnyanesh T, Bengel FM. Cardiac sympathetic neuronal imaging using PET. Eur J Nucl Med Mol Imaging 2007;34:S74–85.