Automated Assessment of Alcohol-Induced Impairment of Balance in Male and Female Social Drinkers

Abstract

The acute administration of alcohol reliably impairs the ability to balance when standing. The standardized field sobriety test (SFST) uses alcohol-induced impairment of body stability to indicate probable alcohol intoxication. Given that body sway is used in the detection of alcohol impairment and intoxication, it is surprising that little research with humans has incorporated new technology that provides automated neuromuscular control assessment. Therefore, the purpose of this study was to examine the dose response to the acute effects of alcohol below and at the legal limit for driving in the U.S. on balance impairments, as measured by the Biosway Portable Balance System (Biodex Medical Systems, Inc.). Fourteen social drinkers attended three separate sessions where they received alcohol (0.0, 0.3, and 0.6 g/kg alcohol). Body sway with eyes open and eyes closed was assessed at 45 minutes after dose administration when BrAC was peaking for both active alcohol doses (.04 g% and .08 g%). The results indicated that body sway was significantly increased in the 0.6 g/kg alcohol conditions when compared to the placebo and 0.3 g/kg alcohol conditions. Body sway was not significantly elevated in the 0.3 g/kg alcohol condition compared to placebo. The results from this study suggest that this new technology may be of interest to alcohol researchers and the police as a more precise assessment of balance.

Introduction

The consumption of alcohol reliably impairs the ability to balance while standing. Laboratory research demonstrates that the acute administration of alcohol leads to dose dependent increases in body sway (Ando et al., 2008; Nieschalk et al., 1999; Zoethout et al., 2012). The U.S. standardized field sobriety test (SFST), used by police to identify probable intoxication, relies heavily on observable alcohol-induced motor impairment (National Highway Traffic Safety Administration (NHTSA), 1999). Starting in 1975, the NHTSA sponsored research to develop a method that was standardized for police officers to evaluate probable intoxication in drivers. Since 1981, the SFST battery has been used, involving three tests; one leg stand, walk and turn, and horizontal gaze nystagmus. The one-leg stand and walk and turn tests both require balance as the person who may have consumed alcohol (or other drugs that impair driving) must stand on one leg for 30 seconds or walk nine heel-toe steps on a line, turn, and then walk another nine heel-toe steps while remaining on the line. The horizontal gaze nystagmus test also has a motor control component as it assesses if there is the involuntary jerking of the eyes as the eyes move to the side that is seen with alcohol intoxication. As such, the three components of the NHTSA SFST rely heavily on observing motor impairments. Note that the NHTSA does not involve a cut score where the individual passes or fails. The police officer uses judgement when observing performance on these tests, including whether the person can follow instructions. While used every day in the U.S., the SFST has also not been updated since 1981. Therefore, the purpose of this laboratory-based study with social drinkers was to determine if new technology that can assess balance impairments in medicine is also useful in the detection of alcohol-induced impairment.

The Biodex Portable Balance System (Biodex Medical Systems Inc., Shirley, NY) is commercially available medical equipment used in for both assessment and rehabilitation of various neuromuscular conditions related to cerebellum functioning (El-Gohary et al., 2016; Furtado et al., 2016; Nepocatych et al., 2018; Zesiewicz et al., 2017). The patient steps on a platform and receives feedback and training via a computer screen at eye level. For this study, the clinical test of sensory integration and balance (CTSIB) was used as it is a measure of body sway. For this CTSIB test, the individual is asked to stand as still as possible on a platform for 30 seconds. The test is given when eyes are open and when eyes are closed. Published data from our lab at a peak breath alcohol concentration (BrAC) of approximately .07 g% indicated that body sway was greater under alcohol when compared to placebo (Marczinski et al., 2018). If this technology may be useful to both researchers and in the field, it would be important to know if this finding is replicable and if body sway increases at alcohol doses lower and higher than this .07 g%, including the current .08 g% U.S. legal limit for driving. As such, this study incorporated alcohol doses with a peak of .04 and .08 g% to determine CTSIB body sway in social drinkers with eyes open and eyes closed.

While researchers may be interested in this technology, the practical application of this technology could not be overemphasized. The vast majority of alcohol-involved automobile crashes occur at night and involve young people between 21–24 years old (Fell et al., 2009). Police officers who are tasked with administering the SFST often must do so under unsafe conditions. Dark road-sides and potentially intoxicated individuals can lead to unintentional injuries for both the officer and the individual in question. Further, the SFST has multiple components, takes time to administer, and does not have an obvious criteria for test failure. While it is not the purpose of this study to determine if the Biosway could replace the SFST, this study would be helpful in determining if an automated and very quick test would also be able to detect probable alcohol intoxication.

Therefore, the purpose of this study was to determine if alcohol-induced increases in body sway are detectable by an automated assessment developed to assess neuromuscular control. Using a within-subjects design, social drinkers (n = 14) were recruited for three separate dose administration sessions where they received alcohol (0.0 g/kg, 0.3 g/kg, and 0.6 g/kg). Body sway was measured using the CTSIB Biosway test when eyes were open and when eyes were closed. It was hypothesized that alcohol administration would increase body sway in a dose dependent manner.

Method

Participants

Fourteen adults (7 men and 7 women) between the ages of 21 and 24 years (M = 21.86, SD = 0.86) participated in this study. The self-reported racial makeup of the sample was all white participants. For ethnicity, no participants self-identified as Hispanic. Participants had a mean (SD) weight of 73.45 kg (10.81). Volunteers completed questionnaires that provided demographic information, alcohol use habits, and physical and mental health status.

Exclusion criteria included a self-reported psychiatric disorder, substance use disorder, head trauma, or other CNS injury. In addition, volunteers with a score of the 5 or higher on the Short Michigan Alcoholism Screening Test (Selzer et al., 1975) and/or a score of 8 or higher on the Alcohol Use Disorders Identification Test (Barbor et al., 1989) were also excluded. Furthermore, individuals who did not regularly consume alcohol (i.e., fewer than 2 standard drinks per month) were excluded. All volunteers provided informed consent before participating. The Northern Kentucky University Institutional Review Board approved this study, and volunteers received $200 for their participation in the entire four session study.

Apparatus and Materials

Personal Drinking Habits Questionnaire (PDHQ: Vogel-Sprott, 1992).

The PDHQ measures an individual’s current, typical alcohol use habits including: (a) number of standard drinks (i.e., bottles of beer, glasses of wine, and shots of liquor) typically consumed during a single drinking occasion, (b) dose (grams of absolute alcohol per kilogram of body weight typically consumed during a single drinking occasion), (c) weekly frequency of drinking, and (d) hourly duration of a typical drinking occasion. The PDHQ also measures previous experience with alcohol in terms of the number of months that an individual has been drinking on a regular basis or customarily on social occasions.

Timeline Follow-Back (TLFB: Sobell & Sobell, 1992).

The TLFB assesses self-reported daily patterns of alcohol consumption during the past 30 days including maximum number of continuous days of drinking, maximum number of continuous days of abstinence, total number of drinking days, total number of drinks consumed in the past month, highest number of drinks consumed in one day, total number of heavy drinking (5+ drinks) days, and total number of ‘drunk’ days (i.e., days on which the participant self-reported feeling intoxicated).

Biosway Balance Assessment.

The Biosway Portable Balance System (Biodex Medical Systems, Inc., Shirley, NY) was used to assess balance using a standardized clinical test of sensory integration and balance (CTSIB) testing protocol (eyes open and eyes closed) that is provided by the manufacturer. For this test, the participant stood on the firm surfaced platform and looked at a color touchscreen display at eye level. After demographic information including body height was entered by the research assistant into the touchscreen display, the participant places his/her feet about hip width apart on the platform in the standardized location marked on the platform, with feet slightly turned out at 10 degree angle. The participant was instructed to look forward at the screen during all testing and not look at the feet or anywhere else in the testing room.

For the CTSIB test protocol (eyes open first and then eyes closed second), the participant was given a three second audible beep countdown before testing commenced. For each 30 second trial, the participant is instructed to keep as still as possible and keep his or her body centered (i.e., not learning forward, backward, or to either side). There were two trials with eyes open followed by two trials with eyes closed. The trials were separated by a 10-second break. The Biosway calculated an overall sway index as the standard deviation of the recorded position away from the center in degrees for the three trials for each test (eyes open and then eyes closed). Higher sway index scores indicated greater deviations from the center (i.e., greater variability in postural stability indicated more body sway). The entire testing including the proper positioning of the feet takes about 5 minutes to complete.

Procedure

Pre-laboratory Screening.

Individuals who responded to the advertisements contacted the research assistant by e-mail to set up a time to participate in a telephone intake-screening interview. During the telephone interview, volunteers were informed that the purpose of the experiment was to study the effects of alcohol on behavioral and mental functioning. Volunteers were told that they would be asked to come to the lab for four sessions to perform computerized tasks and complete paper-and-pencil questionnaires. Moreover, they were informed that they would receive a beverage to consume on all sessions except the first one, and that the beverage they would receive on each session could contain the maximum dose of alcohol found in 4 beers. The research assistant determined if the participant met all eligibility requirements to participate. Before any test session, participants were required to fast for 2 hours, abstain from any form of caffeine for 8 hours, and abstain from alcohol for 24 hours. Eligible subjects then made an appointment for one baseline session and three dose administration sessions.

Baseline Session.

Each participant was tested individually by a trained research assistant. Testing occurred between the hours of 10:00 a.m. and 4:00 p.m. Testing times within one subject were kept as similar as possible and did not vary more than 4 hours from the baseline session start time. Upon arrival in the lab for the baseline session, the participant was asked to provide informed consent. The participant also completed a general health questionnaire, PDHQ, and TLFB questionnaires. The participant was weighed. Finally, the participant practiced the Biosway tasks (eyes open and eyes closed).

Dose Administration Sessions.

Sessions two through four were dose administration sessions. At the start of every dose administration, the participant was weighed and completed a medical screening questionnaire to ensure that the participant was healthy and had not recently taken any prescribed or over-the-counter medications. A zero breath alcohol concentration (BrAC) was confirmed using an Intoxilyzer Model 400 (CMI Inc., Owensboro, KY). The participant provided a urine sample in a private bathroom in the lab and the research assistant immediately tested the urine sample for drug metabolites and pregnancy (women only). For drug metabolites, the testing was for benzodiazepines, barbiturates, tetrahydrocannabinol, cocaine, amphetamines, and opiates (uVera Diagnostics, Inc., Norfolk, VA). No participants had a positive drug test on any session.

On each test day, the participant received one of three possible doses of alcohol (0.60 g/kg, 0.30 g/kg, or placebo). The doses were prepared using (1) 1.82 ml/kg vodka + 3.64 ml/kg decaffeinated soft drink for the 0.60 g/kg dose condition, (2) 0.91 ml/kg vodka + 1.82 ml/kg decaffeinated soft drink for the 0.30 g/kg condition, and (3) 1.82 ml/kg decaffeinated soft drink for the placebo condition. Doses were calculated by body weight and the alcohol doses were reduced by 87% for female participants to avoid sex differences in obtained BrACs. For the placebo, 10 ml of vodka was floated on the surface of the beverage to give the drink an alcoholic scent (Marczinski et al., 2011). For the vodka, 40% alcohol/volume Smirnoff Red Label vodka, No. 21 (Smirnoff Co., Norwalk, CT) was administered and Squirt (Dr. Pepper Snapple Group, Plano, TX) was chosen for the decaffeinated carbonated soft drink. Dose administration was single blind as the research assistant was monitoring breath alcohol concentrations (BrACs). Dose order was counterbalanced between subjects. These doses were chosen as the 0.60 g/kg dose would result in a moderate peak BrAC of approximately .08 g% (Marczinski et al., 2011) which is the U.S. legal limit for driving.

Participants were given their beverage in a plastic cup and were asked to consume the drink within 10 minutes. Drinking was self-paced. Participants were informed that they might receive alcohol, a decaffeinated soft drink, or a combination of these during all of the test sessions. The exact contents of the beverages were never disclosed to the participants. After dose administration, participants relaxed. BrACs were measured at 40, 60, and 80 min. after drinking. Participants also provided breath samples at those times during sessions where no alcohol was administered.

After the BrAC at 40 min. after drinking began, participants completed the Biosway task. Upon completion of the testing period, participants waited in the laboratory until their BrAC fell below .02 g%., at which time they were released. Participants who had not received alcohol were immediately released after the testing battery concluded. All participants were paid and debriefed following the completion of the final session.

Data Analyses

For the dependent measures obtained, the data were submitted to separate within-subjects analyses of variance (ANOVAs) with Dose (0.0 g/kg, 0.3 g/kg, and 0.6 g/kg) as the within subjects factor. For significant results, pairwise comparison of means utilized a Bonferroni correction. Sex was included as an initial between-subjects factor for all initial analyses but no main effects or interactions with sex were obtained for the measures after dose administration. Given the small sample size, sex is only reported for the analyses of the drinking habits and BrACs where sex differences could be anticipated. The alpha level was set at .05 for all statistical tests and SPSS 24 was used to conduct all analyses.

Results

Demographic Characteristics and Self-reported Alcohol Use

Table 1 provides all demographic and baseline questionnaire measures for the male and female participants. Possible sex differences in these baseline measures were tested using two tailed independent samples t tests. Results revealed that males weighed significantly more than females, t(12) = 2.33, p = .038. No other significant differences on any of the other measures were observed (ps > .059).

Table 1.

Demographic characteristics and alcohol use for the male (n = 7) and female (n = 7) participants.

	Males		Females
	M	SD	M	SD
Age	22.29	0.95	21.43	0.54
Weight (kg)	79.26	12.10	67.63	5.27
Body Mass Index	24.67	4.24	25.91	3.97
SMAST	0.00	0.00	0.00	0.00
AUDIT	5.43	1.82	4.71	2.06
Personal Drinking Habits
Questionnaire (PDHQ):
History (months)	59.71	20.01	46.00	17.89
Frequency (occasions/week)	1.32	0.66	1.50	0.82
Alcohol dose (g/kg)	1.11	0.61	0.81	0.39
Duration (hour)	4.43	3.30	3.43	1.57
Timeline Follow-back (TLFB):
Continuous drinking days	1.57	0.79	2.14	1.57
Continuous abstinence days	6.86	2.48	9.14	2.85
Total no. drinking days	5.71	1.50	6.29	3.40
Total no. drinks	28.00	20.46	22.57	18.06
Highest no. drinks in 1 day	7.71	4.15	6.71	4.39
Heavy drinking days	2.86	2.04	1.43	1.52
Drunk days	1.86	1.77	2.43	2.23

Breath Alcohol Concentrations

No detectable BrACs were observed under the 0.0 g/kg placebo condition. For all participants, mean (SD) BrACs obtained at the 40, 60, and 80 min. readings were .08 g%(.02), .07 g% (.01), and .06 g% (.01) for the 0.60 g/kg dose condition, and .04 g% (.01), .03 g% (.01), and .02 g% (.01) for the 0.30 g/kg dose condition. BrACs were examined by a 2 (Alcohol Dose: 0.3 v. 0.6 g/kg) × 3 (Time: 40, 60, v. 80 min.) × 2 (Sex) mixed design ANOVA, where Alcohol Dose and Time were treated as within-subjects factors and Sex was treated as a between-subjects factor. A main effect of Dose, F(1,24) = 96.17, p < .001, η² = .889, was obtained, as BrACs were higher in the 0.6 g/kg dose condition when compared to the 0.3 g/kg dose condition. A main effect of Time, F(2,24) = 45.32, p < .001, η² = .791, was obtained, as BrACs were highest at 40 min., and declined at 60 min., and declined further at 80 min. Pairwise comparisons with a Bonferroni correction indicated that the three mean BrACs taken at 40, 60, and 80 min., were all significantly different from one another, ps < .001. The Biosway assessment occurred after the 40 min. BrAC reading and therefore was performed at approximately .04 g% and .08 g% for the two dose conditions. There was no significant main effect of Sex and no significant interactions with Sex for the analysis of BrACs, ps > .588.

Biosway CTSIB Balance Assessment

Figure 1a reports the mean (SD) values for the mean sway index scores in degrees for the Biosway CTSIB eyes open test. The results of a within-subjects ANOVA revealed a main effect of Dose, F(2,26) = 12.53, p < .001, η² = .491. As shown in the Figure and confirmed with pairwise comparisons with a Bonferroni adjustment for multiple comparisons, greater body sway was observed in the 0.6 g/kg alcohol dose condition when compared to both the 0.0 g/kg and 0.3 g/kg conditions, ps < .004. The 0.0 g/kg and 0.3 g/kg conditions did not differ.

Figures 1a and 1b.

Mean sway index (degrees) for all alcohol dose conditions when participants were instructed to keep their eyes open (1a) and eyes closed (1b). Vertical bars indicate the standard error of the mean.

Figure 1b reports the mean (SD) values for the mean sway index scores in degrees for the Biosway CTSIB eyes closed test. The results of a within-subjects ANOVA revealed a main effect of Dose, F(2,26) = 8.21, p = .002, η² = .387. As shown in the Figure and confirmed with pairwise comparisons with a Bonferroni adjustment for multiple comparisons, greater body sway was observed in the 0.6 g/kg alcohol dose condition when compared to both the 0.0 g/kg and 0.3 g/kg conditions, ps < .024. The 0.0 g/kg and 0.3 g/kg conditions did not differ.

Discussion

The current study examined how the acute administration of low and moderate doses alcohol would result in increased body sway, as measured by the automated Biosway CTSIB tests with eyes open and eyes closed. The results indicated that body sway was significantly increased in the 0.6 g/kg alcohol conditions when compared to the placebo and 0.3 g/kg alcohol conditions for both eyes open and eyes closed tests. Body sway was not significantly elevated in the 0.3 g/kg alcohol condition compared to placebo. In sum, the results from this study suggest that the Biosway CTSIB tests may offer a technological advance for those interested in a more precise assessment of balance after the acute administration of alcohol.

The results presented have public health implications or at least may spur more research on this topic that could be eventually beneficial to public health. The above findings suggest that the speed and safety of administration of the SFST could be improved with an automated assessment that is sensitive to the impairing effects of alcohol at doses relevant to the legal limit of driving in the U.S. The results for this study indicated clear increases in body sway at the .08 g% BrAC level. By contrast, body sway was not appreciably different than placebo below the legal limit. It is known that law enforcement activities are very effective in reducing alcohol impaired driving crashes (Fell et al., 2014). However, only 3% of drivers are exposed to sobriety checkpoints. This may not be surprising given that the SFST in its current form takes time to administer (i.e., more than 15 minutes for the SFST) and the determination of impairment lacks a clear failing score given test administration procedures. If the SFST was faster and automated, testing may more frequently administered by police officers which could ultimately improve public safety.

This study has limitations that should be noted and may offer guidance for the planning of future research studies. First, this study included a small and non-diverse social drinking sample (i.e., all white social drinking college students) who were participating in a laboratory study and were not highly motivated to minimize behavioral impairment. While the highest rates of impaired driving are seen in the 21–25 age range, which was very similar to the participants in this study, not all drinking drivers are similar to the sample in this study. Future research should incorporate a more diverse sample including individuals of different ages, races, ethnicities, length of drinking history, and current drinking habits. Importantly, a double-blind study that compares the effect size and dose sensitivity for both the SFST and this Biosway assessment would be a logical next step in research. It would also be important to show that impairment was detectable when BrAC is declining, as this is when individuals tend to drive. Despite the current limitations, the current results offer some reassurance that technological improvements can improve methodological approaches to human psychopharmacology which may ultimately improve public health.

Public Health Significance.

This laboratory study demonstrated that an automated assessment of body sway can detect balance impairments at alcohol doses at the U.S. legal limit for driving. As balance impairments are relied upon as indicative of probable intoxication in the standardized field sobriety test, the use of this technology could increase both the speed and safety of administration of this test.

Likewise, this technology may be of interest to laboratory researchers interested in how the acute and chronic effects of drug administration alter gross motor control.