Changes in heart rate variability in patients of alcohol dependence syndrome—Do we have a biomarker for craving?

Abstract

Background:

Heart rate variability (HRV) is a measure of physiological function of the autonomic nervous system. HRV decreases immediately after alcohol consumption. In those with alcohol dependence, HRV is lower than in healthy individuals even after several days of abstinence.

Aim:

The study aimed to investigate whether abstinence from alcohol had an association with changes in HRV.

Methods:

An observational study conducted at the psychiatry unit of a tertiary care hospital. 100 patients of age 18–45 years fulfilling criteria for alcohol dependence syndrome were included along with 100 healthy age and gender matched controls. HRV indices (pNN50, SDNN, rMSSD, HFHRV, LFHRV, and LF/HF ratio) were compared, before and after 4 weeks of detoxification with de-addiction treatment as usual.

Results:

At admission, compared to the controls, cases had significantly lower means for the HRV indices. Four weeks after detoxification, cases had significantly higher values of HRV indices. Craving scores correlated with changes in HRV indices. No association was observed with age of onset of drinking or dependence, family history, or choice of anti-craving medication.

Conclusion:

HRV was reduced in cases of ADS; however, with detoxification and continued abstinence, it increased within a period of 4 weeks. Also, craving was associated with change in HRV indices. Thus, HRV is affected by alcohol use but it reverses with alcohol detoxification and abstinence. Hence, can be used as a reliable biomarker of abstinence as well as craving.

INTRODUCTION

There is a growing body of literature showing that chronic and heavy alcohol use has toxic effects on the autonomic nervous system (ANS) which leads to cardiac autonomic neuropathy.[1] The heart rate and heart rate variability (HRV), are regulated through the ANS, and its dysregulation reflects impairment of the system’s ability to cope with varied environmental challenges. Quantification of HRV through ECG recordings is divided into frequency and time domain methods, the earlier being more common, using time series analyses to evaluate HRV within different spectral components, and the latter having a high-frequency (HF-HRV) and a low-frequency component (LF-HRV). LF-HRV is determined by both sympathetic nervous system (SNS) and parasympathetic nervous system (PNS) inputs to the heart, while HF-HRV is a more specific measure of cardiac PNS activity. The time domain methods evaluate variability through indices, including, proportion of interbeat intervals differing by 50ms (pNN550), standard deviation of all interbeat intervals (SDNN), and square root of the mean squared difference of successive interbeat intervals (rMSSD), which correlate well with the high-frequency component of HRV. Reduced HRV has been shown to be related to risk factors for cardiovascular disease, so autonomic dysfunction could be a mediator of the association of cardiovascular risk factors with CVD.[2] Garland et al.[3] noted that alcohol-dependent patients who relapsed, exhibited a significantly greater HFHRV reactivity to stress-primed alcohol cues than patients who did not relapse. Participants who relapsed exhibited greater HFHRV reactivity to alcohol cues than those who did not relapse by the 6-month follow-up.[3] Rechlin et al.[4] reported reductions in HRV in patients with alcohol dependence. A review by Romanowicz et al.[5] suggested association of acute alcohol ingestion with HRV. In a meta-analysis Quintana et al.[6] noted that alcohol dependence was associated with low HRV. Ina an Indian study, Ganesha et al.[7] noted that sleep HRV of alcohol dependent patients had lower SDNN, pNN50, and RMSSD than healthy controls. In another study by Quintana et al.,[8] the authors found that HRV accounted for around 12.1% of the craving in patients with alcohol dependence after controlling for confounding factors.

It seems that HRV is a reliable marker of autonomic dysfunction in patients with alcohol dependence, however, the research on HRV in alcohol dependence is still in its infancy and the results are discrepant. Analysis of the literature show that there is a gap in the evidence of effects of alcohol on HRV and its improvement with time. Thus, this study was carried out to understand the dynamics of heart rate variability among patients with alcohol dependence syndrome vis a vis controls and changes, if any, with treatment.

METHODOLOGY

The study was carried out at a de-addiction Centre of a Tertiary Care Hospital in Delhi. The study design was an observational study with a control arm to establish the normative values for HRV indices in a healthy sample. Cases included patients admitted at the de-addiction centre of a tertiary care hospital for alcohol dependence syndrome (ADS) for in-patient de-addiction treatment during the period July 2018 and July 2020. Attendants of the admitted patients (but not relatives or family members), age and gender matched, were taken up as controls to establish a baseline for comparison of HRV indices. Age matching was carried out by selecting controls within a range of ± 2 yrs compared to the cases. Resting heart rate variability analyses was carried out at admission, at one week and at four weeks post admission, in 100 in-patients who were suffering from ADS (ICD 10) and in 100 age and gender matched healthy individuals who were the attendants of these patients. The sample size was calculated based on the study of Suhas Ganesha, et al.[7] who observed that SDNN, RMSSD, and PNN50 in cases was 76.6 (19.4), 23.7 (10.9), and 2.8 (4.2) respectively and in controls was 96.8 (24.2), 34.5 (9.1), and 6.6 (4), respectively. Taking these values as reference, the minimum required sample size keeping the power of study at 80% and level of significance at 5% is 57 patients in each study group. To reduce margin of error and considering attrition rate, total sample size taken was 200 (100 patients per group).

Inclusion criteria for cases

1. All male patients in the age group 18–45 years undergoing in-patient treatment for Alcohol Dependence Syndrome.

2. Patients who gave consent for participation.

Inclusion criteria for controls

Age and gender matched healthy individuals with no lifetime h/o alcohol use disorder or cardiovascular disease.

Exclusion criteria for cases

1. Patients who did not/could not consent to the study.

2. Patients with psychiatric disorders other than ADS.

3. Patients with serious medical illness including hypertension and cardiac disease.

Exclusion criteria for controls

1. Individuals who did not consent to the study.

2. Participants with lifetime psychiatric disorders

3. Participants with serious medical illness including hypertension and cardiac disease.

Socio-demographic data was collected as per semi-structured format of the centre. Socio-economic status was assessed using the Modified Kuppuswamy scale. The details about the duration of alcohol consumption and quantity of alcohol consumed was recorded. The tools used for assessment were:

1. ICD Symptom Check List, Psychoactive Substance Use Syndromes Module, 2004 was used for the diagnosis of alcohol dependence syndrome

2. Mini-International Neuropsychiatric Interview (MINI 6.0) was used as a screening tool for other psychiatric disorders.

3. Clinical Institute Withdrawal Assessment of Alcohol Scale, Revised (CIWA-Ar) was used for assessment of alcohol withdrawal.

4. Obsessive compulsive drinking scale (OCDS) was used for assessment of craving

5. Severity of alcohol dependence questionnaire (SADQ) was used to assess the severity of alcohol dependence syndrome.

CIWA-Ar was measured at admission (CIWA_t1). OCDS was administered at t1 and after 4 weeks of admission (t2). Resting heart rate variability was measured by BioPack machine (MP36/35 Four Channel Data Acquisition System with inbuilt software AcqKnowledge 4.4.1) for 10 minutes in the supine position. After the measurement of resting HRV, all the ADS patients underwent four to six weeks of treatment with detoxification, group therapy, individual counselling, and anticraving medication. The treatment included: Inj Thiamine 100mg im for 5 days. If in withdrawal state, detoxification with benzodiazepines for a period of 5–7 days.

Post detoxification, anti-craving medication was prescribed—naltrexone 50 mg OD or acamprosate 333 mg 2 TDS. Baclofen, in doses of 30–60 mg/day was prescribed for patients with deranged LFT. Cap Fluoxetine 20 mg/day was used as anti-craving medication for negative affective drinking, if depressive symptoms were present. The patients were also treated with group therapy, individual therapy or other psychological therapies as indicated. Among the cases, HRV was measured at admission (t1), after detoxification (t2) and after 4 weeks of detoxification (t3).

Among the controls, HRV was measured at a single occasion when they accompanied the patient, and the same values were used for comparison with cases at various time points. The following Indices were measured:

1. Proportion of NN intervals exceeding 50ms (pNN50).

2. Standard deviation of all NN intervals (SDNN).

3. The square root of the mean of the sum of squares of differences between adjacent NN intervals (RMSSD).

4. Low frequency power (LF).

5. High frequency power (HF).

6. LF/HF Ratio.

HRV Data was analyzed using the Bio Pack software.

Outcome measures

1. Obsessive compulsive drinking scale scores at t1 and t3.

2. HRV indices.

Statistical analysis

Data was tested for normality. Scores for OCDS and HRV indices were noted to be positively skewed and were log transformed to reduce skewness. Both categorical and continuous variables were tabulated. Appropriate parametric tests, including repeated measures ANOVA and repeated measures correlation were applied for inferential statistics. The data was entered in MS EXCEL spreadsheet and analysis was done using Statistical Package for Social Sciences (SPSS) version 23.0 and R version 4. A P value of < 0.05 was considered statistically significant.

Ethical considerations

Ethical clearance was obtained from the Institutional Ethical Committee, prior to the start of the study. Written informed consent was taken from all study participants. Confidentiality and privacy were ensured at all stages.

RESULTS

The cases and controls were similar with respect to age, gender, weight, BMI, education, and socio-economic status [Table 1]. The mean period of detoxification was 4.7 ± 1.2 days. Alcohol use data has been summarized in Table 1.

Table 1.

Demographic variables

Variable	Overall (n=200) Mean±SD	Cases (n=100) Mean±SD	Controls (n=100) Mean±SD	Test Statistic (t test)	P
Age (years)	36.0±4.81	36.0±4.81	35.9±4.83	-0.28	0.78
Height (cm)	172.84±5.04	172.43±5.23	173.26±4.84	1.16	0.25
Weight (kg)	72.75±7.98	72.13±8.39	73.37±7.55	1.10	0.27
BMI (kg/m2)	24.5±2.58	24.4±2.60	24.5±2.57	0.04	0.97
Education (years)
10th	16 (16%)	22 (22%)	38 (19%)	1.449*	0.485
12th	65 (65%)	63 (63%)	128 (64%)
BA	19 (19%)	15 (15%)	34 (17%)
Income (Rs)	37770±2784.72	37630±2754.81	37700±2763.73	0.357	0.721
Socio-economic status	Lower Middle	Lower Middle	Lower Middle	-	-
Age at Onset of drinking	-	19.8±2.78	-	-	-
Age at dependence	-	25.9±2.92	-	-	-
Years of drinking	-	16.2±4.63	-	-	-
Years of Dependence	-	10.1±4.35	-	-	-
SADQ	-	25.85±6.37	-	-	-
Age at onset ≥18 yrs	-	31	-	-	-
Age at dependence ≥25 yrs of age	-	45	-	-	-

SADQ – Severity of Alcohol Dependence Questionnaire

Comparison of HRV indices between cases and controls HRV indices were compared between the cases and the controls and the difference was statistically significant for all the indices at the time of admission except LF/HF [Table 2].

Table 2.

Comparison of HRV indices at admission between cases and controls

HRV index	Controls (n=100)	Cases (n=100)	t Test Statistic (df=98)	P
pNN50	14.3±17.29	4.1±8.66	5.78	<0.001
SDNN	34.2±14.52	29.4±13.62	2.64	0.017
RMSSD	31.7±18.38	23.1±13.69	3.94	<0.001
LF	398.3±347.16	235.0±337.18	4.16	<0.001
HF	335.9±342.79	167.2±289.28	5.21	<0.001
LF/HF	2.3±2.83	2.2±2.00	-1.04	0.30

HRV – Heart rate variability, pNN50 – proportion of NN50 exceeding 50ms, SDNN – standard deviation of all interbeat intervals, rMSSD – square root of the mean squared difference of successive interbeat intervals, LF – Low frequency component of HRV, HF – High frequency component of HRV, LF/HF – Ratio of Low and High frequency components

HRV indices within the cases- Though the HRV indices within the cases increased significantly over the 3 time points, we tried to find out whether the cases themselves formed heterogeneous groups when the ages of onset of drinking, age of onset of dependence, family history of AUD, and choice of anti-craving medication were considered. A mixed RMANOVA with age of onset of drinking (Age_EO) above and below 18 yrs of age as groups revealed that the HRV indices pNN50, RMSSD [Figure 1], and HF differed significantly between the groups [Table 3]. However, no difference in HRV indices was seen when the groups were divided based on age of onset of dependence, family history, and choice of anti-craving medication.

Figure 1.

Profile Plot to represent the RMSSD vs time interaction (time points t1, t2, t3) among the cases. X axis – Time points (1 - at admission, 2 – post detoxification, 3 – after 4 weeks of detoxification), Y Axis – Estimated marginal means of rMSSD, rMSSD - square root of the mean squared difference of successive interbeat intervals

Table 3.

Age of onset of drinking (Age_EO) vs HRV indices interaction between time points at admission, after detoxification and at 4 weeks of admission (*represents interaction)

Univariate Tests
Source	Measure	df	F	Sig (P)	Partial Eta Squared η2
HRV	pNN50	1.84	57.10	<0.001	0.37
	SDNN	1.54	47.13	<0.001	0.33
	RMSSD	1.93	75.08	<0.001	0.43
	LF	1.74	46.45	<0.001	0.32
	HF	1.85	62.90	<0.001	0.39
	LF_HF	1.68	0.25	0.74	0.00
HRV* Age_EO	pNN50	1.84	4.41	0.02	0.04
	SDNN	1.54	1.56	0.22	0.02
	RMSSD	1.93	3.41	0.04	0.03
	LF	1.74	2.70	0.08	0.03
	HF	1.85	3.13	0.04	0.03
	LF_HF	1.68	0.58	0.53	0.01

HRV – Heart rate variability, pNN50 - proportion of NN50 exceeding 50ms, SDNN - standard deviation of all interbeat intervals, rMSSD - square root of the mean squared difference of successive interbeat intervals, LF – Low frequency component of HRV, HF – High frequency component of HRV, LF_HF – Ratio of Low and High frequency components, df – degrees of freedom, F – F statistic, Sig – Statistical Significance

Craving measures (OCDS scores)

OCDS scores were measured at admission and after 4 weeks of de-addiction treatment. It was noted that the OCDS scores reduced significantly over the period of 4 weeks. We also measured the association of OCDS scores with HRV indices at various time points. A two-way RMANOVA revealed that there was a significant interaction between OCDS scores and RMSSD and HF scores [Table 4] [Figure 2]. A repeated measures correlation of OCDS with HF and RMSSD revealed a significant negative correlation between the OCDS scores and the HRV indices [Figure 3].

Table 4.

Association between OCDS scores and RMSSD between time points, at admission and 4 weeks after admission (*represents interaction)

General Linear Model
Test of Within Subjects
Source	df	F (1,99)	Sig (P)	Partial Eta Squared η2
OCDS	1	87.08	<0.001	0.47
RMSSD	1	172.37	<0.001	0.64
OCDS* RMSSD	1	172.16	<0.001	0.64

OCDS – Obsessive Compulsive Drinking Scale, rMSSD - square root of the mean squared difference of successive interbeat intervals, df – degrees of freedom, F – F statistic, Sig – Statistical Significance

Figure 2.

Association between OCDS scores and RMSSD before (1) and after detoxification (2) at 4 weeks for the cases. X axis – (1 - OCDS scores before detoxification, 2 - OCDS after detoxification), Y Axis – Estimated marginal means of rMSSD, OCDS – Score on Obsessive Compulsive Drinking Scale, rMSSD – square root of the mean squared difference of successive interbeat intervals

Figure 3.

Repeated measures correlation between OCDS scores and RMSSD indices before after detoxification for the cases, X axis – OCDS scores, Y Axis –rMSSD scores, OCDS – Obsessive Compulsive Drinking Scale, rMSSD – square root of the mean squared difference of successive interbeat intervals

Effect of Anti-craving medication

We also tried to delineate whether anti-craving medications differed in the effects on craving and HRV indices. It may be noted that there was no medication*OCDS score interaction indicating that the choice of anti-craving medication did not have an effect on the reduction in OCDS scores meaning they were all equally effective or they had no effect at all.

DISCUSSION

In this observational study, we included 100 patients of alcohol dependence syndrome being treated at a tertiary care centre in Delhi. The controls were age, gender, and socio-economic status matched attendants of the patients that were unrelated to the patients. These cases were followed up for a period of 4 weeks during which they stayed at the de-addiction centre and were given treatment as usual in the form of anti-craving medication, motivational enhancement therapy, group therapy, and family therapy. The treatment provided was structured and uniform for all the patients.

As shown in Table 1, the cases and the controls were not significantly different with respect to their age, BMI, education, income, and socio-economic status. These cases and controls did not suffer from any cardiac illness that could influence the results of heart rate variability.

In a study by Rechlin et al.[4] in 1996 it was demonstrated that HR-HRV was significantly lower in people drinking alcohol for 15 or more years as compared to the age-matched controls. A number of studies have shown that HRV indices are lowered in patients with alcohol dependence syndrome compared to healthy controls.[9,10,11,12,13,14,15]

After one week of detoxification the HRV indices among the cases increased so that the cases and controls were not significantly different with respect to the time and domain measures, except pNN50 which was still lower than that in controls. After 4 weeks of admission, all the HRV indices showed an increase among the cases and the SDNN and rMSSD indices overshot the values of the controls in that now they had significantly higher values than the controls. In our result, both the time and frequency domain measure showed a linear increase with time during the hospitalization.

Though the HRV indices within the cases increased significantly over the points (admission, one week, and four weeks), we tried to find out whether the cases themselves formed heterogenous groups when the ages of onset of drinking, age of onset of dependence, family history of AUD, and choice of anti-craving medications were considered. It was seen that when the age of onset of drinking above and below 18 years of age were taken as groups, there was a significant HRV* group interaction with respect to pNN50, rMSSD, and HF-HRV. This implies that the patients with late onset of drinking recovered faster with a higher increase in HRV indices compared to those with early onset drinking. However, no difference in HRV indices was seen when the groups were divided based on the age of onset of dependence, family history and choice of anti-craving medication.

Studies have been done in the past that have evaluated the relationship between HRV and craving or relapse. Quintana et al.[8] in 2013 found that the resting HF-HRV was negatively related to craving even after controlling for drinking, anxiety, and age. This study demonstrated that habitual, moderate drinkers of alcohol showed increased levels of HF-HRV relative to nonhabitual drinkers. Also that the alcohol consumption groups did not differ by factors known to influence HRV like physical activity, sex distribution and BMI. These findings of this study were similar to the results seen in our study when the relation of HRV with basic characteristics was studied.

In our study, OCDS scores were measured at the time of admission and after four weeks of de-addiction treatment. It was observed that the OCDS scores reduced significantly over the period of four weeks. We also tried to find an association of HRV measures with reduction in craving and fond a significant interaction between the OCDS scores and rMSSD and HF-HRV scores. This means that there is a significant negative correlation between the OCDS scores and HRV indices and possibly there is a common factor that is leading to a change in both the variables. Craving (OCDS scores) showed significant negative correlation with both rMSSD and HF-HRV).

In a clinical review on the anti-craving therapy for alcohol use disorder by Shen et al.,[16] the author talked about prescribing the mentioned drugs because benefits of these anti-craving drugs are far outweighing their possible side effects, or the harms of untreated AUD itself. In our study, no difference was found among the various anti-craving medications prescribed on the OCDS scores or the HRV indices.

Reduced heart rate variability (HRV), a marker of autonomic dysfunction, has been shown to be associated with an increased risk of incident myocardial infarction, cardiovascular mortality, and death from other causes in general populations, and to be associated with a poor prognosis of cardiovascular diseases (CVD). Reduced HRV has been shown to be related to risk factors for cardiovascular disease, so autonomic dysfunction could be a mediator of the association of cardiovascular risk factors with CVD. Q10An increase in HRV has an impact on health and well-being. Increased HRV is associated with indices of psychological well-being such as cheerfulness and calmness, motivation for social engagement and psychological flexibility.[17] This well-being may result from the reduction craving during the period of hospitalization with the various kinds of treatment provided to these patients.

Thus, it can be inferred from the foregoing discussion that measurement of HRV at admission and subsequently during the period of hospitalization or in the OPD setting may provide an insight into the patient’s craving and the improvement with the treatment. Patients who do not respond to treatment and those who continue to have high craving will not show changes in HRV and thus HRV, particularly rMSSD and HF-HRV may be considered as biomarkers of improvement in the condition of the patients. Confounding variables, such as age, sociodemographic background, psychopathology, and intellectual ability, may be associated with HRV, independent of the problem of alcohol use disorder. There was an effort to match subjects and controls based on these confounding factors, however, lacunae may remain. Comorbidities associated with heavy alcohol intake such as depression, tobacco use, sleep problems, and medications used for their treatment may also affect HRV, though we tried to exclude the patients who had significant cardiac morbidity The study was carried out among in-patients and only for a short term of 4 weeks. Thus, craving in the real-life scenario could not be measured. Also, relapses in these patients over a period of 6 months to 1 yr could have provided better insights into the changes in HRV with craving and subsequent alcohol use. The HRV measures for controls were made only once that may be a limitation to the study as the measures may change over time. Also, though the CIWA-Ar scores were insignificant at t1 (i.e.) first measurement point of HRV), it could have altered the HRV indices, that may be considered as a limitation. There were a few more confounding factors such as anxiety, stress, and sleep, that are known to impact HRV but were not systematically accounted for (though psychiatric disorders were ruled out by MINI). To strengthen HRV’s validity as a craving biomarker, future studies should incorporate longitudinal control group data, covariate adjustments for withdrawal severity and psychiatric symptoms, and dynamic HRV responses to alcohol cues. Additionally, longer-term follow-up, including relapse tracking, would help establish HRV as a predictor of treatment outcomes.

CONCLUSION

In summary, our study concludes that HRV is reduced in ADS patients. Although the results suggest that vagal activity reduction is associated with chronicity and alcohol dependence, HRV is reversible with the treatment. ANS noticeably adapts in response to alcohol use pattern and plays vital role in craving and future relapses of illness. HF-HRV and RMSSD appear to be sensitive markers of these changes. Hence, HRV may be a useful biomarker for assessment of effect of treatment and a measure of craving in patients of alcohol use disorder.

Conflicts of interest

There are no conflicts of interest.

Funding Statement

Nil.