Common Hand Sanitizer May Distort Readings of Breathalyzer Tests in the Absence of Acute Intoxication


  • Syed S. Ali,

    Corresponding author
    1. The Department of Emergency Medicine, Behavioral Emergencies Research Lab, UC San Diego, San Diego, CA
    • University of California San Diego, San Diego, CA
    Search for more papers by this author
  • Michael P. Wilson MD, PhD,

    1. the Department of Emergency Medicine, UC San Diego Health System, San Diego, CA
    2. The Department of Emergency Medicine, Behavioral Emergencies Research Lab, UC San Diego, San Diego, CA
    Search for more papers by this author
  • Edward M. Castillo PhD, MPH,

    1. the Department of Emergency Medicine, UC San Diego Health System, San Diego, CA
    2. The Department of Emergency Medicine, Behavioral Emergencies Research Lab, UC San Diego, San Diego, CA
    Search for more papers by this author
  • Peter Witucki MD,

    1. the Department of Emergency Medicine, UC San Diego Health System, San Diego, CA
    Search for more papers by this author
  • Todd T. Simmons MD,

    1. the Department of Emergency Medicine, UC San Diego Health System, San Diego, CA
    Search for more papers by this author
  • Gary M. Vilke MD

    1. the Department of Emergency Medicine, UC San Diego Health System, San Diego, CA
    2. The Department of Emergency Medicine, Behavioral Emergencies Research Lab, UC San Diego, San Diego, CA
    Search for more papers by this author

  • Presented at the Society for Academic Emergency Medicine annual meeting, Chicago, IL, May 2012.
  • The authors have no relevant financial information or potential conflicts of interest to disclose.

Address for correspondence and reprints: Syed S. Ali; e-mail:



The use of alcohol-based hand sanitizers has recently become widespread. To the authors' knowledge, no previous study has examined whether application of ethanol-based hand sanitizers by the person operating a common breathalyzer machine will affect the accuracy of the readings. This was a prospective study investigating whether the use of hand sanitizer applied according to manufacturer's recommendations (Group I), applied improperly at standard doses (Group II), or applied improperly at high doses (Group III) had an effect on breathalyzer readings of individuals who had not ingested alcohol.


The participants of the prospective study were divided into three groups to assess the effect of hand sanitizer on breathalyzer readings. Group I used one pump (1.5 mL) of hand sanitizer (Purell), allowing the hands to dry per manufacturer's recommendations; Group II used one pump (1.5 mL), without allowing the hands to dry; and Group III used two pumps (3 mL), without allowing the hands to dry. Breathalyzer measures for each group are presented as medians with interquartile ranges (IQR) and ranges. Differences between each sequential group (I vs. II and II vs. III) were assessed using a Mann-Whitney U-test (p < 0.05 significant).


There were 25 study participants in each group for a total of 75 participants. The initial breathalyzer readings of all study participants were 0.000 g/dL. The median breathalyzer reading was 0.004 g/dL in Group I (IQR = 0.001 to 0.008 g/dL), 0.051 g/dL in Group II (IQR = 0.043 to 0.064 g/dL), and 0.119 g/dL in Group III (IQR = 0.089 to 0.134 g/dL). Measures between each subsequent group were all statistically different (p < 0.001).


The use of common alcohol-based hand sanitizer may cause false-positive readings with a standard hospital breathalyzer when the operator uses the hand sanitizer correctly. The breathalyzer readings are further elevated if more sanitizer is used or if it is not allowed to dry appropriately.


Los Antisépticos Comunes para las Manos Pueden Distorsionar las Lecturas de los Alcoholímetros en Ausencia de Intoxicación Aguda


El uso de antisépticos para las manos a base de alcohol ha llegado a ser universal. Según el conocimiento de los autores, ningún estudio previo ha examinado si la aplicación de antisépticos para manos a base de alcohol por la persona que manipula una máquina de test de alcoholemia afectará a la certeza de las lecturas. Este estudio prospectivo investigó si el uso de antiséptico para manos aplicado según las recomendaciones del fabricante (Grupo I), aplicado incorrectamente a las dosis estándar (Grupo II) o aplicado incorrectamente a dosis altas (Grupo III) tuvo un efecto en las lecturas del alcoholímetro en aquéllos que no habían ingerido alcohol.


Los participantes del estudio prospectivo se dividieron en tres grupos para valorar el efecto del antiséptico de manos en las lecturas del alcoholímetro. El Grupo I utilizó un pulso (1,5 ml) de antiséptico para manos (Purell) y permitió secarse las manos según las recomendaciones del fabricante; el Grupo II utilizó un pulso (1,5 ml) de Purell, sin permitir secarse las manos; y el Grupo III utilizó 2 pulsos (3 ml) de Purell sin permitir secarse las manos. Las medidas del alcoholímetro para cada grupo se presentaron como medianas con rangos intercuartílicos (RIC), y rangos. Las diferencias entre cada grupo secuencial (I vs II y II vs III) se realizaron mediante el test de la U de Mann-Whitney (p significativa < 0,05).


Hubo 25 participantes en cada grupo del estudio, con un total de 75 participantes. La lectura inicial del alcoholímetro fue de 0.000 g/dL. La mediana de la lectura del alcoholímetro fue de 0,004 g/dL en el Grupo I, 0,051 g/dL en el Grupo II y 0,119 g/dL en el Grupo III. Las medidas entre cada grupo subsiguiente fueron todas estadísticamente significativas (p < 0,001).


El uso de un antiséptico común para manos puede causar lecturas falsamente positivas con un alcoholímetro estándar hospitalario cuando el operador usa el antiséptico para las manos de forma correcta. Las lecturas del alcoholímetro serán cada vez más altas a mayor dosis de antiséptico o si no se permite el secado de forma adecuada.

Breathalyzer machines are commonly used to make a determination of the degree of alcohol intoxication in emergency department (ED) patients, as bedside clinical skills are surprisingly poor in this regard.[1, 2] Breathalyzer machines are also often used by law enforcement to make a preliminary determination of intoxication, with a common legal threshold of 0.080 g/dL.[1-3] Given the importance of both clinical and legal determinations of alcohol use, previous work has focused on the distortion of breathalyzer readings by food or drink ingested by the subject, such as an alcohol-based gum or mouthwash.[4-6] Subjects must therefore be supervised for a minimum of 15 minutes before breathalyzer testing, during which time the subject must not have ingested alcoholic beverages or other fluids, regurgitated, vomited, eaten, or smoked. Subjects may also undergo serum alcohol testing, in which the use of alcohol and other volatile disinfectants to cleanse the skin are discouraged.[7]

Whether or not hand sanitizers also cause a distortion of breathalyzer readings is unclear. Most commercially available hand sanitizers, such as Purell, Germ-X, and Labo, contain ethyl alcohol, which is thought to be necessary to denature microbes. Other previous studies have documented that these commonly available hand sanitizers do not significantly elevate the blood alcohol levels in individuals who have applied it to themselves.[8-12] In one study, for instance, five health professionals with initial blood ethanol levels below 0.005 g/dL applied 5 mL of a Kimberly-Clark hand-sanitizer product containing 62% denatured ethyl alcohol to their hands. They rubbed the hand sanitizer over both hands until they were dry and then repeated this procedure 50 times over a period of 4 hours. The study revealed that the blood ethanol level upon completion of the 50 applications was still less than 0.005 g/dL in all five cases. They concluded that no clinically significant transdermal absorption of ethanol occurred in the participants.[13] Another study by Brown et al.[9] assessed cutaneous ethanol and isopropanol absorption after intensive (30 times per hour) use of alcohol-based sanitizer solutions by health care workers. The results indicated that ethanol was detectable in the breath (0.001 g/dL to 0.0025 g/dL) 1 to 2 minutes after exposure and in the serum in two of the 20 health care workers 5 to 7 minutes after exposure.9 Additional studies confirmed these levels, suggesting that excessive application of ethanol-based hand sanitizer does not produce a clinically significant effect on blood alcohol levels or breathalyzer results.[8, 11-13]

It is unclear, however, whether a hand-sanitizer mixture applied to a person measuring breathalyzer levels could cause falsely elevate readings of the person whose alcohol level is being measured. There is reason to suspect that it might, since alcohol is volatile and thus might cause distortion of the breathalyzer readings. The objective of this study was therefore to determine whether the application of alcohol-based hand sanitizers on the hands of the person holding a breathalyzer affected readings of others.


Study Design

This was a prospective trial in which each participant served as his or her own control. The local institutional review board approved the entire study, including sample size calculations, prior to initiation of the project. All participants completed the informed consent process.

Study Setting and Population

The study was conducted in the ED of the University of California, San Diego Hillcrest Medical Center. Potential study participants included patients and visitors in the ED who had not consumed alcohol in the previous 24 hours. Exclusion criteria included: 1) people less than 18 years old, 2) patients identified as being too acutely ill to participate in the study, 3) people unable to speak English, 4) employees at work such as staff or medical students, and 5) people unable to participate in the informed consent process. Further identifying information about patient demographics, other than age, was not collected.

Study Protocol

The breathalyzer, an Alco-Sensor III breathalyzer device (Intoximeters Inc., St. Louis, MO), was calibrated per manufacturer's recommendations before each reading was taken. Participants were instructed to blow into the breathalyzer, which is commonly used at local hospitals. Participants who had a result > 0 were excluded from the study. Immediately afterward, the experimenter applied hand sanitizer to his or her hands. The hand sanitizer used in this trial was commercially available Purell (62% ethyl alcohol, 38% inactive ingredients including a small amount of isopropyl alcohol; GOJO Industries, Inc., Akron, OH). In Group I, this was one pump (1.5 mL) of hand sanitizer with rubbing the hands briskly until dry per manufacturer recommendations; in Group II, this was one pump (1.5 mL) of hand sanitizer without drying; and in Group III, this was two pumps (3.0 mL) of hand sanitizer without drying. A second breathalyzer reading was then immediately obtained. As there are no estimates of how long the hands should dry, participants were tested both immediately (assumed to represent improper but common practice), and after an arbitrary length of time for drying, in this case, when the experimenter noted that his or her hands were dry. The participants were instructed to blow into the machine for approximately 2 seconds before the measurement button was pressed.

The Alco-Sensor III breathalyzer device uses fuel cell rather than semiconductor technology to detect and measure breath alcohol content. Fuel cell technology involves using an electrode to break down the alcohol molecules in a breath sample into electrons, protons, and acetic acid. This produces an electric current in the process. Measuring the strength of the current allows the device to indicate the quantity of alcohol in the sample.[14] Fuel cell breathalyzer devices such as the Alco-Sensor III are considered accurate and reliable and have been tested and approved by the U.S. Department of Transportation. Alco-Sensor III has been approved by the National Highway Traffic Safety Administration for use as an evidential breath measurement device.[15]

Since there is no extant literature on which to base an estimate of effect size, a pilot study among the authors of the study (MW and TS) was conducted on themselves. Based on these pilot data, effect size was estimated at 0.07 with a standard deviation (SD) of 0.05. Thus, a sample size calculation predicted that at least 10 participants would need to be enrolled to have a power of slightly greater than 95%. Since the first group used half as much hand sanitizer as the third group, at least 20 participants were predicted to be needed for this group and therefore was the target for each group to keep group size equal.

Data Analysis

The outcome measure for this study was the second breathalyzer reading after hand sanitizer was applied. The medians and interquartile ranges (IQRs) are presented for each group. Differences between each sequential group (I vs. II and II vs. III) were compared using the Mann-Whitney U-test. A p-value < 0.05 was considered statistically significant. All statistical analyses were conducted using the IBM SPSS Statistics 19.0 software package (SPSS, Inc., Chicago, IL).


Twenty-five individuals participated in the first group, which was more than needed based on sample size calculations. Subsequently, the same numbers of individuals were enrolled in the second and third groups to keep the numbers in each group equal. Thus, a total of 75 individuals participated in the experiment. There were 35 females, and the average age of all participants was 34 years (range = 20 to 59 years, SD ± 10 years). All 75 participants had initial breathalyzer readings of 0. No potential subjects were excluded based on a positive initial breathalyzer reading.

The median breathalyzer readings and IQRs after the application of hand sanitizer are reported in Figure 1. After one pump was applied according to manufacturer's recommendations with drying (Group I), breathalyzer measurements ranged from 0.000 to 0.019 g/dL, with a median of 0.004 (IQR = 0.001 to 0.008). Five participants had no detectable alcohol. After one pump of hand sanitizer was applied without drying (Group II), breathalyzer measurements ranged from 0.020 to 0.109 g/dL, with a median of 0.051 (IQR = 0.043 to 0.064). After two pumps of hand sanitizer were applied without drying (Group III), breathalyzer measurements ranged from 0.020 to 0.166, with a median of 0.119 g/dL (IQR = 0.089 to 0.134). Notably, this is a little more than twice the result obtained with Group II and is above the legal intoxication threshold in most states of 0.08 g/dL. Breathalyzer measures between Group I and Group II, as well between Group II and Group III, were significantly different (all p < 0.001).

Figure 1.

Median breathalyzer readings and IQRs after the application of hand sanitizer. IQR = interquartile range.


The use of hand sanitizers is common in EDs, as medical staff are advised to apply hand sanitizer before and after contact with each patient. The results of this study indicate that some common alcohol-based hand sanitizers, such as Purell, may affect breathalyzer readings, presumably by vaporization of the hand sanitizer on the part of the person holding the breathalyzer machine. This affects breathalyzer readings in patients who have not ingested alcohol. The degree of distortion is larger if hand sanitizer is applied incorrectly, but exists even with its correct application.

Results from this study regarding the application of ethanol-based hand sanitizer and its effect on breathalyzer readings may raise potential concerns. In clinical settings such as the ED, intentional ingestions of alcohol-based hand sanitizers have increased in frequency.[16] Even in patients who have not ingested alcohol, as in the study here, breathalyzer machines must provide precise results in cases where an alcohol level is desired because they may affect a clinician's ability to properly assess a patient. In addition, patients improperly assumed to be intoxicated might be held longer than necessary in the ED resulting in crowding and/or longer wait times for other patients. Another potential concern outside the hospital is that many states use breathalyzer machines to administer preliminary breath tests. Accuracy is critical because police officers may rely on the breathalyzer reading to establish probable cause for a driving under the influence arrest.[17]


A potential limitation of this study is that only one brand of hand sanitizer was used, and the results may not generalize across all other brands, each of which may contain different amounts of active ingredients. This is likely of lesser concern, as most health care hand sanitizer brands contain >60% alcohol.[16] Another potential limitation is that only one breathalyzer machine was used. Different models may be more or less susceptible to this type of distortion.


The use of an alcohol-based hand sanitizer can cause false-positive readings of a breathalyzer when the operator uses the hand sanitizer correctly. The breathalyzer readings are further elevated if more sanitizer is used or if it is not allowed to dry appropriately. Health care workers and others who rely on these machines to make judgments about the alcohol level of others should be careful not to use hand sanitizer immediately before operating a breathalyzer machine and to use it according to manufacturer's recommendations when they do. Future experiments should be conducted to determine whether the use of gloves diminishes the effect of ethanol-based hand sanitizer on breathalyzer readings.