Association between Circulating Adiponectin and Heart Rate Recovery in Women with Polycystic Ovarian Syndrome

ABSTRACT Context Adiponectin (APN) is reported to be correlated closely with autonomic nervous function in different clinical settings. Heart rate recovery (HRR) is a noninvasive and readily obtainable indicator, which reflects the coordinated interplay between parasympathetic reactivation and sympathetic withdrawal. Objective This study aimed to investigate the relationship between serum APN and HRR in polycystic ovarian syndrome (PCOS) women. Design Eighty-nine PCOS women were enrolled and divided into two groups. Women with HRR values slower than 12 beats were defined as Blunted HRR Group. APN levels were compared between Blunted HRR Group and Normal HRR Group. Multivariate logistic regression analysis and multiple linear regression analysis were performed to determine which clinical variables were independently associated with HRR and APN levels, respectively. Results Twenty-three women were categorized into Blunted HRR Group, in which APN level was significantly lower than Normal HRR Group. Age, body mass index, hypertension, and APN were independent factors of attenuated HRR in PCOS women. Meanwhile, multiple linear regression analysis showed age, dyslipidemia, and homeostasis model assessment-insulin resistance (HOMA-IR) were closely associated with APN levels in PCOS women. Conclusions Our findings suggested that decreased APN concentration was closely associated with HRR blunt in PCOS women. Further studies are needed to explore the underlying interactions between APN and autonomic nervous function.


Introduction
Polycystic ovary syndrome (PCOS) is one of the most common endocrine disorders of reproductive age, affecting 15-20% of women. 1 The 'Rotterdam' diagnostic criteria defines that women must present with two of the following three symptoms/signs -hyperandrogenism, chronic anovulation/oligomenorrhea and polycystic ovaries. 2 Epidemiological literature documents that women with PCOS have an increased prevalence of cardiometabolic complications, such as dyslipidemia, hypertension, type 2 diabetes and metabolic syndrome. 3 Although the exact etiology of PCOS has not been elucidated yet, imbalance of autonomic nervous system is confirmed as a key contributor for reproductive disorders and cardiometabolic outcomes by accumulating clinical studies. 4 Heart rate recovery (HRR), defined as the difference between heart rate (HR) at peak exercise and HR one minute after exercise, reflects the dynamic balance and coordinated interplay between parasympathetic reactivation and sympathetic withdrawal. 5 Meta-analysis has proved that slow HRR has the prognostic value of predicting cardiovascular events and mortality in different populations, including asymptomatic adults, patients with coronary artery diseases and patients with heart failure. 5 The significance of HRR, a noninvasive and readily obtainable indicator, was verified repeatedly in those investigations, thus we hypothesized that HRR evaluation in PCOS women would be of clinical interest, and it would be meaningful to explore the factors which have close associations with HRR, such as adipokines.
Adiponectin (APN), a 30 kDa circulating serum adipokine, is predominantly secreted by adipose tissue and is down-regulated in states of obesity and insulin resistance. 6 A recent meta-analysis by Li et al., 7 which included 1,944 PCOS women and 1,654 healthy women from 38 studies, reported lower APN levels in PCOS women compared with controls. Moreover, a growing number of investigations indicated that APN was related to autonomic nervous function in adults and children. [8][9][10] Recently, an observation carried out in Australia reported that APN was inversely associated with sympathetic activity in PCOS women. 11 However, there were fewer similar studies focusing on the association between APN and autonomic nervous imbalance in Chinese PCOS women. In this respect, we designed this study to determine whether APN level was independently correlated with HRR in PCOS women, which would be helpful to explore the interactions between adipokines and autonomic nervous activities in PCOS.

Subjects
This study was approved by Shanghai Minhang Hospital Ethics Committee. Written consent was obtained from each participant before information was collected. During the period from January 2016 to November 2020, 89 women who were diagnosed as PCOS according to the Rotterdam ESHRE-ASRMsponsored PCOS criteria 2 from the Department of Gynecology and Obstetrics were enrolled in this study. Participants were excluded if they had any of the following conditions: pregnancy, malignancy, liver or renal failure, severe orthopedic problems or severe muscular/ skeletal disorders that would prohibit exercise test, acute or chronic infectious diseases, taking chronotropic drugs (i.e., ß-blockers, amiodarone, or calcium-channelblockers) or sympathomimetic drugs within 12 weeks.

Clinical Data Collection
Clinical data including medical history, exercise habit (taking more than 3 times/week and 30 minutes/time regular exercise was defined as having exercise habit), smoking habit and alcohol habit were collected from each participant. Body mass index (BMI) was calculated as the ratio of weight divided by height squared (kg/m 2 ). Systolic blood pressure (SBP rest ) at rest, diastolic blood pressure (DBP rest ) at rest and HR at rest (HR rest ) were also recorded.

HRR Measurement
Every participant underwent an incremental cardiopulmonary exercise test (CPET) on a treadmill. CPET was performed according to a symptom-limited Bruce's protocol, with continuous electrocardiographic monitoring. Blood pressure (BP) was measured and recorded regularly. The test was terminated for any of the following reasons: the exertion sore greater than 17 (Borg scale); actual HR more than 90% of age predicted maximum HR; SBP more than 200 mmHg; typical chest discomfort; severe arrhythmias; more than 1 mm of horizontal or downsloping ST segment depression. After achieving peak exercise, the test was almost immediately stopped. Therefore, the subjects in our CPET did not have a 'cool-down' period. HRR was obtained by subtracting HR at the first minute of recovery from peak HR obtained during exercise. It was reported that the value of HRR may be affected by the small workload during the 'cool-down' period, decreasing its diagnostic sensitivity. 12,13 The whole process was conducted by a cardiologist and a physician.
Substantial cohort data verified that attenuated HRR was an independent predictor for all-cause and cardiovascular mortality, although adopting different cutoff points of HRR. The value of less than 12 beats was widely used as the cutoff in those investigations, including patients referred for exercise test, 14 candidates for angiography, 15 survivors of acute myocardial infarction 16 and heart failure patients. 17 Therefore, we defined women whose HRR values were slower than 12 beats in our study as Blunted HRR Group. Then, 89 PCOS women was divided into two subgroups: Blunted HRR Group (n = 23) and Normal HRR Group (n = 66).

Statistical Analysis
The Kolmogorov-Smirnov normality test was used to determine if the data was normally distributed. Descriptive statistics were presented as the mean ± SD for variables with normal distribution, median values and range for variables with non-normal distribution or percentage for categorical variables. Variables with non-normal distribution were log-transformed to achieve normal distribution before analysis. Differences between two groups were assessed with two-sided Fisher exact tests, chi-square tests for categorical variables and Student t-tests for continuous variables. Multivariate logistic regression analysis was performed to determine which clinical variables were independently associated with blunted HRR and the results were reported as odds ratios (ORs) and 95% confidence intervals (95% CIs). Meanwhile, multiple linear regression analysis was used to examine the effect of clinical variables on APN levels and the results were reported as β coefficients and 95% CIs. All statistical analyses were performed using the software package SPSS, version 21.0 (SPSS Inc., Chicago, USA). P< .05 was considered significant statistically.

Results
The demographic and clinical characteristics of 89 PCOS women were presented in Table 1. Twentythree patients were categorized into Blunted HRR Group. There was no woman having the habit of smoking or drinking alcohol in either group. Age, BMI, PCOS duration and the prevalence of hypertension history were greater in Blunted HRR Group than those of Normal HRR Group (P < .05), while the ratio of women having exercise habit was lower in Blunted HRR Group than that of Normal HRR Group (P< .01). As for SBP rest , DBP rest , HR rest , prevalence of coronary artery disease, prevalence of dyslipidemia and prevalence of glucose metabolism disorder, there was no significant difference in two groups (P> .05). In regard to biochemical indexes, there was no significant difference in FBG, IN, HOMA-IR, AST, ALT, TC, HDL-C, LDL-C, TG, Cr, FSH, LH, TT, and DHEAS (P > .05). However, APN level in Blunted HRR Group was significantly lower than that of Normal HRR Group (P< .05).
Moreover, those PCOS women with hypertension had lower APN levels than those without hypertension (P < .05, Table 3). There were also statistically significant differences in APN levels in dyslipidemia subgroup and glucose metabolism disorder subgroup (P < .05, Table 3).

Discussion
In this study, we found that PCOS women with blunted HRR had lower values of circulating APN than those women with normal HRR. Furthermore, in multivariate analysis, age, BMI, hypertension and APN concentration were independent risk factors of HRR impairment.
PCOS women were characterized as elevated androgen levels, anovulation and menstrual irregularity. Substantial evidences prove that sympathetic nervous system hyperactivity is associated with those vital symptoms. 4 Moreover, accumulating clinical data demonstrate that PCOS women have increased risks of cardiovascular diseases and metabolic disorders, compared with age-matched healthy women. Meanwhile, increased muscle sympathetic nerve activity, increased levels of adrenergic metabolites in the serum and urine and lower HR variability were reported as different manifestations of generalized increase in sympathetic tone in PCOS women. 18 HRR, the quantification of HR decrease after exercise, is a straightforward method and a highly reproducible tool in assessing autonomic nervous function. Extensive investigations have recognized that it is not only a powerful factor predicting all-cause mortality, but also a potential marker predicting health outcomes including cardiovascular diseases. 5 It was reported that attenuated HRR was a precursor to hyperglycemia as well as an indicator of cardiovascular dysfunction. [19][20][21] In the past decades, HRR has been gradually evaluated in clinical research of PCOS women. Giallauria F et al. found that HRR was lower in young PCOS women compared with healthy subjects (12.9 ± 1.8 beats vs. 20.4 ± 3.1 beats). 22 Similarly, Kaya C et al. found HRR was decreased in PCOS women compared with age-and BMI-matched women (15.4 ± 1.9 beats vs. 24.2 ± 3.4 beats). 23 Additionally, it was reported that 89 out of 243 young PCOS patients without known risk factors for cardiovascular diseases presented with abnormal HRR, in which study abnormal HRR was defined as lower than 18 beats for standard exercise testing. 24 In the present study, our definition of blunted HRR was lower than 12 beats, and 23 out of 89 women exhibited blunted HRR. Previous extensive data supports that postexercise HRR includes two distinct phases: a fast phase, characterized by abrupt decay of HR and determined by parasympathetic reactivation, and a slow phase, characterized by gradual decay of HR and predominantly determined by sympathetic withdrawal. HRR one minute after exercise is considered mostly to be in the period of fast phase. 5,25 However, we couldn't tell the exact content of parasympathetic or sympathetic tone in the HRR values in the current study. We adopted it as an index representing autonomic balance as well as sympathovagal interplay so that we investigated the relationship between APN and the balance marker. Although we found the lower APN associated with blunted HRR, the observation result wasn't powerful statistically enough to provide extension to the existing literature on the sympathetic arm of autonomic dysfunction in PCOS women.
Because HRR possesses such important predictive values, researchers made efforts to explore its influencing factors. Previous studies showed age was one of those factors, and our multivariate analysis result confirmed it. The prevalence of abnormal HRR was reported to be greater in older than younger individuals. 26,27 This age-effect was also supported by Buchheit et al., 28 whose study revealed HRR in children were higher compared with adolescents and adults. Variables with normal distribution were presented as mean ± SD; Variables* with non-normal distribution were presented as median with range, which were log-transformed to achieve normal distribution before analysis. Physical fitness and training may also influence HRR. Several cross-sectional investigations showed that athletes or physically trained individuals had greater HRR than sedentary individuals, 5,29 so we collected the information about exercise habits of those PCOS women at enrollment. Although exercise habit wasn't associated with HRR independently, the ratio of women taking regular exercise in Blunted HRR Group was significantly lower than Normal HRR group. We found increased BMI was another risk factor for blunted HRR in PCOS women. As early as in 2008, it was reported abnormal HRR was inversely correlated with BMI in overweight PCOS women. 22 Then it was found that weight loss in overweight and obese women with PCOS was associated significantly with improvement in HRR. 30 Another positive finding in the current study was hypertension was also an independent determinant of attenuated HRR in PCOS women. A recent research based in China reported that HRR was lower in hypertensive patients than non-hypertensive patients, and it was even lower in hypertensive patients with uncontrolled BP than hypertensive patients with controlled BP. 31 Moreover, a study enrolling 1,855 participants who were healthy and normotensive initially indicated that slow HRR was independently associated with the development of hypertension after average 4-year's follow-up. 32 Hypertension, as a common coexisting condition in PCOS, has complicated interaction with HRR. We deemed that longitudinal HRR assessment in PCOS women was helpful for clinical management both in prehypertension phase and different grades of hypertension.
Besides these above-discussed clinical parameters, some circulating factors were found to be closely correlated with HRR in PCOS, such as C-reactive protein, 23,24 white blood cells count, 24 and homocysteine. 23 Furthermore, adipokines, which are produced by adipose tissue, have been found to mediate the crosstalk between metabolic function and autonomic nervous function in different populations during the past two decades' research. 33 APN, which is the most abundant adipose-released cytokine, not only plays an important role in wholebody energy homeostasis, but also results in protection of the vasculature, heart, lung, and kidney because of its anti-apoptotic properties. 33 In addition to those effects, it was reported that higher APN levels were associated with a more favorable development of cardiovascular autonomic function in individuals with type 2 diabetes. 8 Another clinical observation reported that lower APN levels were related to sympathetic activation in obstructive sleep apnea patients. 34 On the other side, Łukasz Nowak et al. found that blockade of sympathetic nervous system activity increased plasma APN concentration in patients with essential hypertension. 35 Recently, it was shown that increased sympathetic activity was associated with high-molecular-weight APN levels in premenopausal women with PCOS. 11 Our study was a clinical observation, without exploring any underlying mechanism, but those data gave us a hint that HRR, a window to the status of autonomic nervous function, combined with APN measurement might be promising monitoring indicators for stratified management of PCOS women in the future.  Although our study showed a significant difference of APN levels between Blunted HRR Group and Normal HRR Group, it was found HRR wasn't an independent factor of APN concentration after multiple linear regression analysis. However, age, dyslipidemia and HOMA-IR were correlated with APN levels independently. Prior studies have revealed that serum APN levels rise with increasing age and an elevation of ~1 μg/mL of circulating APN for every 10 years of age, in healthy individuals. 6 Meanwhile, the significant relationship between dyslipidemia and APN was consistent with former reports. It was found serum APN was negatively related to TC, LDL-C and TG concentrations, and positively related to HDL-C concentration in nondiabetic subjects. 36 Although the mechanism that accounts for the relation between APN and serum lipids is not fully understood, it was revealed that APN was directly associated with enzymes that regulate lipid metabolism, and those associations were independent of age, sex, BMI, insulin resistance, and systemic inflammation. 36 Additionally, HOMA-IR was found to be significantly associated with APN level in our work. Animal experiments and clinical studies have confirmed APN is strongly involved with PCOS physiopathology and insulin resistance, and some literature proposed that APN might be a new surrogate biomarker to facilitate and improve the determination of IR in this specific female population. 37 Despite those interesting findings in the current study, certain limitations should be considered. First, this was a single-center study in which only a small number of PCOS patients were available for analysis; Second, our work was cross-sectional, it would be more informative to have serial values of APN and HRR to explore causality; Third, there were some factors which were noted to influence serum APN levels, such as smoking habit 38 and moderate alcohol consumption. 39 There was no woman having smoking habit or alcohol habit in our study. Current evidence about the effect of dietary fatty acids on APN values is equivocal, 39 and we didn't collect the information of those women's oral daily supplements, like fish oil or conjugated linoleic acid, which might have potential effect on circulating APN values; Fourth, plenty of previous studies demonstrated HMW-APN associated with insulin sensitivity closely. 40 In addition, it was even identified the importance of HMW-APN/total APN ratio when assessing its relationship with insulin sensitivity. 40 In our study, we chose the index of total APN because most investigations focusing the relationship between APN and sympathovagal balance (such as heart rate variation analysis [8][9][10] and baroreflex sensitivity assessment 41 ) tested total APN concentration. It would have been better if we had collected both HMW-APN and total APN values at first design so that we could have more analysis.

Conclusions
In conclusion, our study demonstrated that APN levels were significantly lower in Blunted HRR Group than Normal HRR Group. Furthermore, in multivariate analysis, age, BMI, hypertension, and APN levels were independent predictors of HRR blunt in PCOS. Further large-scale studies are needed to explore the underlying pathways of interactions between APN and autonomic nervous activities in PCOS women.

Data Availability
All data/results generated during this study are included in this published article or in the data repositories listed in References. All datasets analyzed during the current study are not publicly available but are available from the corresponding author on reasonable request.