J Am Acad Audiol 25:261-267 (2014)
The Influence of Caffeine on Calorics and Cervical Vestibular Evoked Myogenic Potentials (cVEMPs) DOI: 10.3766/jaaa.25.3.5 Kathleen McNerney* Mary Lou Coad* Robert Burkard*t
Abstract Background: Prior to undergoing vestibular function testing, it is not uncommon for clinicians to request that patients abstain from caffeine 24 hr prior to the administration of the tests. However, there is little evidence that caffeine affects vestibular function. Purpose: To evaluate whether the results from two tests commonly used in a clinical setting to assess vestibular function (i.e., calorics and the cervical vestibular evoked myogenic potential [cVEMP]) are affected by caffeine. Research Design: Subjects were tested with and without consuming a moderate amount of caffeine prior to undergoing calorics and cVEMPs. Study Sample: Thirty young healthy controls (mean = 23.28 yr; females = 21). Subjects were excluded if they reported any history of vestibular/balance impairment. Data Collection and Analysis: The Variotherm Plus Caloric Irrigator was used to administer the water, while the l-Portal VNG software was used to collect and analyze subjects’ eye movements. The TECA Evoked Potential System was used for the cVEMP stimulus presentation as well as for the data collection. During cVEMP collection, subjects were asked to monitor their sternocleidomastoid muscle contraction with a Delsys EMG monitor. IBM SPSS Statistics 20 was used to statistically analyze the results via paired f-tests. Results: Analysis of the data revealed that ingestion of caffeine did not significantly influence the results of either test of vestibular function. Conclusions: The results revealed that a moderate amount of caffeine does not have a clinically sig nificant effect on the results from caloric and cVEMP tests in young healthy adults. Future research is necessary to determine whether similar results would be obtained from individuals with a vestibular impairment, as well as older adults. Key Words: Caffeine, calorics, cervical vestibular evoked myogenic potentials (cVEMPs) Abbreviations: C = caffeine session; cVEMP = cervical vestibular evoked myogenic potentials; HSSC = horizontal semicircular canal; NC = no-caffeine session; PSV = peak slow wave velocity; SCM = sternocleidomastoid; UW = unilateral weakness
‘ Department of Rehabilitation Science, University at Buffalo; tDepartment of Otolaryngology, University at Buffalo Kathleen McNerney, PhD, Department of Rehabilitation Science, University at Buffalo, 511 Kimball Tower, Buffalo, NY 14214; Phone; 716-829-6799; Fax: 716-829-3217; E-mail: [email protected] The cVEMP portion of this study was presented at the 2012 annual meeting of the American Balance Society, March 7,2012, Scottsdale, AZ. The portion of the manuscript regarding caloric testing was presented at the annual meeting of the American Balance Society, March 6, 2013, Scottsdale, AZ. The first author (K.M.) received a New Investigator Research Grant from the American Academy of Audiology’s Research Grants in Hearing and Balance Program; this grant program is funded by the American Academy of Audiology Foundation.
SB1
Journal o f the American Academy o f Audiology/Volume 25, Number 3, 2014
atients who complain of dizziness or imbalance are often required to undergo vestibular func tion testing to assess whether their vestibular system has been compromised. Often, these tests include calorics, which assesses the function of the horizontal semicircular canal (HSSC) (Barin, 2008) and may also include cervical vestibular evoked myogenic potentials (cVEMPs), which assesses the function of the saccule (Colebatch et al, 1994; Ferber-Viart et al, 1999; Akin and Murnane, 2008), one of the otolith organs. During calorics, the HSSC is stimulated with warm and cool water, while the eyes are monitored for the presence of nystagmus. In a healthy vestibular system, peak slow wave velocity (PSV) is present and symmetrical between the right and left sides (Barin, 2008). During cVEMP testing, the vestibular system (i.e., the saccule) is acti vated by an auditory stim ulus (i.e., an intense, lowfrequency toneburst or click). The response is most commonly recorded from an electrode placed on a contracted sternocleidom astoid (SCM) muscle and consists of a positive peak (PI) a t approxim ately 13 msec followed by a negative trough (Nl) at approx imately 23 msec (Ferber-Viart et al, 1999; Zhou et al, 2004; Welgampola and Colebatch, 2005). Abnormalities of the cVEMP may include absence of a response, as well as an amplitude asymmetry (i.e., there is a large differ ence [>35%] in amplitude between the right and left sides) (Welgampola and Colebatch, 2001). One of the only published studies th a t evaluated whether caffeine affects the results of tests of vestibular function was published by Felipe et al in 2005. The study evaluated whether the results from calorics, as well as tests of oculomotor function, were different depending on w hether th eir subjects consumed th e ir norm al amount of daily caffeine. The subjects in their study con sisted of 19 females who reported experiencing vestibular symptoms. The study found that the results from these two tests were similar across both test sessions. In a related study, Enriquez et al (2009) found that caffeine did not have any effect on postural stability in healthy young controls when they were asked to stand on a plat form as still as possible for 30 sec with their eyes open and 30 sec with their eyes closed. The purpose of the present study is to investigate if and how caffeine influences the outcome of two common clinical tests of vestibular func tion in healthy young adults, as patients are often asked to abstain from caffeinated foods and beverages 24 hr prior to vestibular testing (e.g., online instructions provided by the following practices: BayCare Clinic, Capital Region Otolaryngology Head and Neck Group, ENTCare).
P
METHODS hirty healthy controls (mean = 23.28, SD = 1.95 yr; 21 females) were tested during two separate sessions. During the caffeine (C) session, subjects
T
SBS
were asked to refrain from caffeine the morning of the test, and when they arrived for the test, they were asked to consume 16 oz. of Starbucks Breakfast Blend coffee, which contained ~300 mg of caffeine immedi ately prior to undergoing caloric and cVEMP testing (McCusker et al, 2003; www.starbucks.com). Subjectswere allowed to add cream and sugar to their coffee if they did not prefer to drink it black. During the no caffeine (NC) session, subjects were asked to refrain from caffeine for 24 hr prior to testing. The two sessions were counterbalanced. Subjects were free of any significant medical history, as well as free of any history of vestibular or balance disorders, and were compensated for their par ticipation. Approval for the study was obtained from the University at Buffalo institutional review board, and writ ten consent was obtained from each of the subjects. When subjects first arrived for the NC session, they were asked to complete a caffeine withdrawal question naire, which allowed us to assess the types of symptoms they were experiencing as a result of not drinking their normal amount of caffeine. The questionnaire assessed a number of symptoms, including headache, tiredness, difficulty concentrating, nausea/vomiting, and anxiety, and was adapted from a study that evaluated the type of caffeine withdrawal symptoms that were experienced by individuals 48 hr after removing caffeine from their diet (Ozsungur et al, 2009). The subjects were asked to rate the symptoms they were experiencing on an 11-point scale; a score of 0 indicated that they were not experienc ing the symptom, while a score of 10 indicated that they would rate that symptom as severe. Subjects were also required to complete a caffeine diary, which allowed eval uation of the average amount of caffeine each of the subjects consumed per week. Subjects were required to include the number and type of beverages, including brand (e.g., Starbucks, Dunkin Donuts, Pepsi, Coke, Red Bull), as well as the total number of ounces they consumed of each beverage each day. Milligrams per week of caffeine consumption was calculated for each subject through the use of the caffeine “dose by product” data reported on www. energyfiend.com, wilstar.com, and www.mayoclinic.com. Initially, subjects underwent tympanometry via the Audio Traveler AA222, which assessed the integrity of the middle ear. This was to ensure that each of the sub jects had normal middle ear function, as even a small con ductive hearing loss (resulting in an air-bone gap of 10 dB or more) can result in a reduced or absent cVEMP to airconducted stimuli (Halmagyi et al, 1994). In addition, tympanometry was performed to assure that the subject did not have a perforated eardrum prior to performing caloric testing. During caloric testing, subjects’ ears were irrigated with warm (44°C) and cool (30°C) w ater via the Variotherm Plus Caloric Irrigator (Atmos, Medizin Tecknik, Lenzkirch, Germany). Eye movement was monitored via in frared goggles. The I-P ortal VNG software
Influence of Caffeine on Calorics and cVEMPs/McNerney et al
(NeuroKinetics, Pittsburgh, PA) was used to collect and analyze the subjects’ eye movements. The percentages of unilateral weakness, directional preponderance, and asymmetry (degrees/second) were recorded and analyzed for each subject, as well as the PSV (degrees/second) for each of the four irrigations (i.e., right warm, left warm, right cool, and left cool). Unilateral weakness values of .007 (incorporating a Bonferroni correction), the only result that approached significance was the left warm irrigation. The Cohen’s d value for this paired comparison was equal to 0.48, which is considered to be a small to medium effect size. The raw data revealed that the PSV was slower during the C session than the NC session, which was also true of the remaining caloric irrigations (see Table 2). Despite the mean PSVs being larger in the NC session than in the C session, the mean directional preponderance and asymmetry results were larger in the C session than the NC session, while the unilateral weakness results were almost identical (i.e., 10.45 [C] vs. 10.47 [NC]) (see Table 2). When evaluating the individual data, there were three subjects who displayed abnormal results on calorics. All three displayed a unilateral weakness (UW). Two sub jects displayed a weakness that persisted throughout both sessions. The results from the two subjects were as follows: S20 (who was classified as a high caffeine
user) during the C session, UW = 43.46%, and during the NC session, UW = 38.50%, and S25 (who was clas sified a moderate user of caffeine) during the C session, UW = 35.95%, and during the NC session, UW = 35.27%. The third subject (S26, who was classified as a low caffeine user) only displayed a weakness dur ing the caffeine session, and it was ju st outside of normal limits (UW = 25.72%). During the NC session, the results were within normal limits for this subject (UW = 22.40%).
cVEMP Figure 1 displays an example of the cVEMP from an individual subject during the C and NC sessions. Table 3 fists the P I and N1 latencies as well as the P1N1 ampli tudes, for the right and left sides, during the C and NC sessions. The results reveal that there were minimal differ ences in mean cVEMP latency and amplitude between the C and the NC sessions. In addition, paired t-tests did not reveal any significant differences between the two ses sions. Clinically, an asymmetry ratio is often calculated, which allows one to determine if there is any difference' in the function of the right and left sides of a patient. Asymmetiy ratios up to 35% are considered to be within normal limits (Welgampola and Colebatch, 2001, 2005). Table 4 fists the asymmetry ratios for all of the subjects for both the C and NC sessions. The present study found mean asymmetry ratios of 15.06% (SD = 14.22) for the C session and 19.33% (SD = 13.33) for the NC session. In addition, statistical analysis of the asymmetry ratios via paired t-tests did not reveal any statistical difference between the C and NC sessions (t(29) = —1.712, p = 0.098, Mainr rence 4.05, S I!difi'ercRce 12.95, d .31). However, when evaluating the individual data, it was found that five subjects had values outside of normal limits. One sub ject (S30) fell outside of normal limits throughout both ses sions (C session: 54.40%; NC session: 49.99%), while the remaining four abnormal asymmetry ratios were as fol lows: S14: 52.81% (C), 25.99% (NC); S16: 37.43% (C), 19.27% (NC); S20: 33.69% (C), 51.53% (NC); S28: 3.41% (C), 44.94% (NC).
DISCUSSION any clinicians ask patients to refrain from con suming caffeine prior to undergoing tests of vestib ular function (ENTCare, Capital Region Otolaryngology Head and Neck Group, BayCare Clinic). The present study revealed that caffeine did not clinically or statisti cally (using a conservative Bonferroni approach) affect the test results between the C and NC sessions. There fore, the authors of the present study conclude that mod erate amounts of caffeine do not substantially affect the clinical interpretation of caloric or cVEMP testing in healthy young adults.
M
SG 4 J
Influence of Caffeine on Calorics and cVEMPs/McNerney et al
Table 2. Mean Peak Slow W ave Velocity (PSV) for Each of the Four Caloric Irrigations, as Well as the Percentage of Unilateral Weakness (UW), Directional Preponderance (DP), and Asymmetry Left cool
Right cool
Left warm
Right warm
DP (%)
UW (%)
Asymmetry (degrees/second)
10.45 8.58 -22.84 20.71 20.97 -20.44 Caffeine 10.47 7.15 -23.92 23.05 21.79 -20.94 No caffeine 0.02 1.43 1.08 2.34 0.82 0.50 Difference .912 .408 .361 .016 .184 .674 P value N o te : The difference row indicates the absolute mean difference between the C and NC sessions. Paired r-test results are row. Following a Bonferroni correction, p s 0.007 was considered to be significant.
Previous studies in our laboratory have also found that caffeine has minimal effects on the sensory orga nization test (SOT) (McNerney et al, forthcoming). Although the study found a statistically significant dif ference between the C and NC sessions for condition 5 (C5) and the composite score, the clinical interpretation of the results was not affected. In addition, the study also evaluated the change in C5 and composite score as a function of average daily caffeine consumption. The results showed that there was a greater change between the sessions for individuals who consumed lower amounts of caffeine as compared to those who consumed higher amounts of caffeine. This would strengthen the argument that (at least for healthy young controls) if individuals are accustomed to drinking moderate amounts of caffeine, asking them to refrain from doing so before undergoing tests of vestibular function is unnecessary. In the present study it is interesting to note that two of the individuals who did not report any history of vertigo or imbalance were found to have a UW, which persisted throughout both sessions. It is important to keep in mind that the variance for calorics is large and the caloric response is dependent on several factors, such as atten tion, the effectiveness of caloric stimulation, and size of the ear canal (Proctor and Glackin, 1985). Therefore, it is not all that improbable that two out of 30 subjects were outside of normal limits. In addition, when we evaluated the cVEMP asymme try ratios, we found that three subjects in the C session NC session
4.18 4.01 0.17 .738 listed in the bottom
revealed an abnormal asymmetry ratio (>35%) and three subjects in the NC session revealed an abnormal asymmetry ratio. Only one of the subjects had an abnor mal asymmetry ratio that persisted across session: (C: 54.4%; NC: 49.99%). Again, this subject did not report any vestibular or balance impairments prior to enrolling in the study. With the other three subjects, because the asymmetry ratios resolved in the opposite session, we can conclude with some certainty that the results are most likely due to the variability th at occurs with cVEMP amplitude. Published research studies have shown that P1N1 amplitude is highly variable and can be influenced by factors such as muscle tension, muscle fatigue, stimulus level, and how the subject is asked to contract the SCM muscle (Akin and Murnane, 2001; Ochi et al, 2001; Isaacson et al, 2006). In the present study, all of the subjects were required to monitor the amount of muscle tension they were producing during the cVEMP recordings, and were required to produce similar amounts of tension during both sessions. How ever, even small differences in the placement of the EMG monitoring electrode may have contributed to the differences in cVEMP amplitude seen across ses sion in these individuals. None of the subjects displayed abnormal results on both caloric and cVEMP tests, with the exception of one subject who displayed an abnormal cVEMP during the NC session; however, this asymmetry ratio resolved during the C session. This finding would be expected, C session
Figure 1. cVEMP recordings from the C and NC sessions from an individual subject. The NC session resulted in an asymmetry ratio of 9.86%, while the C session resulted in an asymmetry ratio of 23.85%.
S 65
J o u r n a l o f th e A m erican A cad em y o f A udiology/Volum e 25, N um ber 3, 2014
Table 3. Mean ± Standard Deviation for the Latency of P1 and N1 as Well as for the P1N1 Amplitude of the VEMP Latency (msec)
Caffeine No caffeine P value
Amplitude (pV)
P1_R
N1_R
P1_L
N1_L
P1N1_R
P1N1_L
13.46 ± 1.1 13.62 ± 1.2 .565
20.68 ± 1.8 20.44 ± 1.6 .471
13.94 ± 1.2 14.13 ± 1.2 .342
20.97 ± 1.8 20.95 ± 1.4 .488
187.63 ± 115.6 206.54 ± 123.4 .963
194.76 ± 148.6 189.72 ± 122.9 .745
Note: R - responses were recorded from the right SCM; L = responses were recorded from the left SCM.
given that the two tests are assessing two separate end organs of the vestibular system. Felipe et al (2005) is one of the only published studies that evaluated whether caffeine affects the results of calorics and oculomotor function. Similar to the present study, they also reported that there were minimal differ ences between the C and NC sessions for a group of 19 women ages 21 to 76 yr (Felipe et al, 2005). With regard to the oculomotor testing, they found that all of the results were normal, regardless of whether subjects were tested with or without caffeine. With regard to caloric testing, they reported a similar number of abnormalities in the C and NC sessions. Restated, if the subjects dis played an abnormal caloric result during the NC session, then they were found to display an abnormal caloric result during the C session. Felipe et al (2005) also reported that individuals expe rienced more anxiety as well as headache from cessation of caffeine. One limitation of the study by Felipe et al (2005) was that everyone in the study received the NC session first. Therefore, individuals could have naturally had more anxiety due to the fact that they were in a new situation and undergoing an unfamiliar test. The present study found that only four out of the 30 subjects indi cated that they were experiencing anxiety during the NC session. Three individuals rated their anxiety level as a one (mild), and one individual rated their anxiety level as a five (moderate). Three out of the four individ uals were in the C session first (particularly the one who rated her anxiety as a five), which means that when they underwent the second (NC) session, they already knew what to expect during each of the tests. Therefore, their anxiety is most likely related to the cessation of caf feine rather than an unfamiliar situation. Headache, on the other hand, was one of the more commonly reported symptoms on our symptom questionnaire. Eleven out of the 30 subjects reported experiencing this symptom as a result of abstaining from caffeine. Subjects rated this symptom from one to four. In addition to anxiety and headache, Felipe et al (2005) also reported that individ uals experienced more nausea and vomiting, as well as more intense vertigo, during the NC session. The present study found that four out of the 30 subjects reported nau sea as a caffeine withdrawal symptom. Two of the sub jects reported a rating of 1, one subject reported a rating of 2, and the final subject reported a rating of 3. Interest ingly, these were the same four subjects who reported
266
that they felt anxious as a result of cessation of caffeine. The conclusions of the present study regarding the caf feine withdrawal questionnaire would have been stron ger if we would have collected another questionnaire on the day that the subjects were undergoing the C session in order to account for symptoms that they may normally experience on a day-to-day basis. CONCLUSIONS
T
he present study evaluated whether the cessation of caffeine is necessary in order to undergo two
Table 4. VEMP Asymmetry Ratios (ARs) for Each of the Subjects during the Caffeine (C) and No-Caffeine (NC) Sessions S2 S3 S4 35 S6 S7 S8 S9 S10 S11 S12 S13 S14 S15 S16 S17 S18 S19 S20 S21 S22 S23 S24 S25 S26 S27 S28 S29 S30 S31
AR_C
AR_NC
0.62 2.19 29.27 12.32 11.20 12.53 5.70 13.14 1.01 10.50 9.02 6.72 52.81* 12.85 37.43* 10.74 11.83 21.37 33.69 0.80 0.94 23.85 21.85 7.23 8.86 24.74 3.41 6.05 54.40* 11.51
8.31 9.08 21.66 13.73 11.05 35.03 7.57 11.02 17.96 13.38 21.24 8.01 25.99 10.51 19.27 1.82 11.47 32.20 51.53* 9.90 1.61 9.86 30.70 19.55 27.55 30.02 44.94* 10.05 49.99* 14.99
*A value that is outside of normal limits.
Influence of Caffeine on Calorics and cVEMPs/McNerney et al
commonly used tests of vestibular function in young healthy adults. The results revealed that the ingestion of a moderate amount of caffeine did not substantially influence the clinical interpretation of either test. This would support the argument that young healthy adults do not need to refrain from drinking caffeine prior to caloric or cVEMP testing. In addition, the results from the caffeine withdrawal questionnaire indicate that the abrupt cessation of caffeine consumption may result in withdrawal symptoms such as headache, nausea, tired ness, and anxiety. In most cases, individuals who are experiencing vestibular symptoms are uncomfortable to begin with. The results from the present study indi cate that we may be causing greater discomfort to our patients than necessary if there are no effects of caf feine on the results from caloric and cVEMP tests. Future studies should focus on whether these same results occur in older adults and in individuals with com promised vestibular function and whether the results are affected by larger amounts of caffeine.
Enriquez A, Sklaar J, Viirre E, Chase B. (2009) Effects of caffeine on postural stability. Int Tinnitus J 15(2):161-163. Felipe L, Simoes LC, Gonjalves DU, Mancini PC. (2005) Evalua tion of the caffeine effect in the vestibular test. Braz J Otorhinolaryngol 71(6):758-762. Ferber-Viart C, Dubreuil C, Duclaux R. (1999) Vestibular evoked myogenic potentials in hum ans: a review. Acta Otolaryngol 119(1):6-15. Halmagyi GM, Colebatch JG, Curthoys IS. (1994) New tests of ves tibular function. Baillieres Clin Neurol 3(3):485—500. Isaacson B, Murphy E, Cohen H. (2006) Does the method of ster nocleidomastoid muscle activation affect the vestibular evoked myogenic potential response? J Vestib Res 16(4-5):187-191. McCusker RR, Goldberger BA, Cone EJ. (2003) Caffeine content of specialty coffees. J Anal Toxicol 27(7):520-522. McNemey KM, Coad ML, Burkard RF. (Forthcoming) The influence of caffeine on the sensory organization test (SOT). J Am Acad Audiol. Nolan S, Heinzer T. (2011) Essentials o f Statistics for the Behav ioral Sciences. 1st ed. New York: Worth Publishers.
REFERENCES
Ochi K, Ohashi T, Nishino H. (2001) Variance of vestibular-evoked myogenic potentials. Laryngoscope lll(3):522-527.
Akin FW, Murnane OD. (2001) Vestibular evoked myogenic poten tials: preliminary report. J Am Acad Audiol 12(9):445-452.
Ozsungur S, Brenner D, El-Sohemy A. (2009) Fourteen welldescribed caffeine w ithdraw al symptoms factor into three clus ters. Psychopharmacology (Berl) 201(4):541-548.
Akin FW, M urnane OD. (2008) Vestibular evoked myogenic potentials. In: Jacobson GP, Shepard NT, eds. Balance Function Assessment and Management. San Diego: Plural Publishing.
Proctor L, Glackin R. (1985) Factors contributing to variability of caloric test scores. Acta Otolaryngol 100(3-4):161-171.
Barin K. (2008) Background and technique of caloric testing. In: Jacobson GP, Shepard NT, eds. Balance Function Assessment and Management. San Diego: Plural Publishing.
Welgampola MS, Colebatch JG. (2001) Vestibulocollic reflexes: normal values and the effect of age. Clin Neurophysiol 112(11): 1971-1979.
Cohen J. (1988) Statistical Power Analysis for the Behavioral Sciences. 2nd ed. Hillsdale, NJ: Lawrence Erlbaum.
Welgampola MSC, Colebatch JG. (2005) Characteristics and clin ical applications of vestibular-evoked myogenic potentials. Neurol ogy 64(10):1682-1688.
Colebatch JG, Halmagyi GM, Skuse NF. (1994) Myogenic poten tials generated by a click-evoked vestibulocollic reflex. J Neurol Neurosurg Psychiatry 57(2):190-197.
Zhou G, Cox LC. (2004) Vestibular evoked myogenic potentials: history and overview. A m J Audiol 13(2): 135-143.
2G7
Copyright of Journal of the American Academy of Audiology is the property of American Academy of Audiology and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.
The Influence of Caffeine on Calorics and Cervical Vestibular Evoked Myogenic Potentials (cVEMPs) DOI: 10.3766/jaaa.25.3.5 Kathleen McNerney* Mary Lou Coad* Robert Burkard*t
Abstract Background: Prior to undergoing vestibular function testing, it is not uncommon for clinicians to request that patients abstain from caffeine 24 hr prior to the administration of the tests. However, there is little evidence that caffeine affects vestibular function. Purpose: To evaluate whether the results from two tests commonly used in a clinical setting to assess vestibular function (i.e., calorics and the cervical vestibular evoked myogenic potential [cVEMP]) are affected by caffeine. Research Design: Subjects were tested with and without consuming a moderate amount of caffeine prior to undergoing calorics and cVEMPs. Study Sample: Thirty young healthy controls (mean = 23.28 yr; females = 21). Subjects were excluded if they reported any history of vestibular/balance impairment. Data Collection and Analysis: The Variotherm Plus Caloric Irrigator was used to administer the water, while the l-Portal VNG software was used to collect and analyze subjects’ eye movements. The TECA Evoked Potential System was used for the cVEMP stimulus presentation as well as for the data collection. During cVEMP collection, subjects were asked to monitor their sternocleidomastoid muscle contraction with a Delsys EMG monitor. IBM SPSS Statistics 20 was used to statistically analyze the results via paired f-tests. Results: Analysis of the data revealed that ingestion of caffeine did not significantly influence the results of either test of vestibular function. Conclusions: The results revealed that a moderate amount of caffeine does not have a clinically sig nificant effect on the results from caloric and cVEMP tests in young healthy adults. Future research is necessary to determine whether similar results would be obtained from individuals with a vestibular impairment, as well as older adults. Key Words: Caffeine, calorics, cervical vestibular evoked myogenic potentials (cVEMPs) Abbreviations: C = caffeine session; cVEMP = cervical vestibular evoked myogenic potentials; HSSC = horizontal semicircular canal; NC = no-caffeine session; PSV = peak slow wave velocity; SCM = sternocleidomastoid; UW = unilateral weakness
‘ Department of Rehabilitation Science, University at Buffalo; tDepartment of Otolaryngology, University at Buffalo Kathleen McNerney, PhD, Department of Rehabilitation Science, University at Buffalo, 511 Kimball Tower, Buffalo, NY 14214; Phone; 716-829-6799; Fax: 716-829-3217; E-mail: [email protected] The cVEMP portion of this study was presented at the 2012 annual meeting of the American Balance Society, March 7,2012, Scottsdale, AZ. The portion of the manuscript regarding caloric testing was presented at the annual meeting of the American Balance Society, March 6, 2013, Scottsdale, AZ. The first author (K.M.) received a New Investigator Research Grant from the American Academy of Audiology’s Research Grants in Hearing and Balance Program; this grant program is funded by the American Academy of Audiology Foundation.
SB1
Journal o f the American Academy o f Audiology/Volume 25, Number 3, 2014
atients who complain of dizziness or imbalance are often required to undergo vestibular func tion testing to assess whether their vestibular system has been compromised. Often, these tests include calorics, which assesses the function of the horizontal semicircular canal (HSSC) (Barin, 2008) and may also include cervical vestibular evoked myogenic potentials (cVEMPs), which assesses the function of the saccule (Colebatch et al, 1994; Ferber-Viart et al, 1999; Akin and Murnane, 2008), one of the otolith organs. During calorics, the HSSC is stimulated with warm and cool water, while the eyes are monitored for the presence of nystagmus. In a healthy vestibular system, peak slow wave velocity (PSV) is present and symmetrical between the right and left sides (Barin, 2008). During cVEMP testing, the vestibular system (i.e., the saccule) is acti vated by an auditory stim ulus (i.e., an intense, lowfrequency toneburst or click). The response is most commonly recorded from an electrode placed on a contracted sternocleidom astoid (SCM) muscle and consists of a positive peak (PI) a t approxim ately 13 msec followed by a negative trough (Nl) at approx imately 23 msec (Ferber-Viart et al, 1999; Zhou et al, 2004; Welgampola and Colebatch, 2005). Abnormalities of the cVEMP may include absence of a response, as well as an amplitude asymmetry (i.e., there is a large differ ence [>35%] in amplitude between the right and left sides) (Welgampola and Colebatch, 2001). One of the only published studies th a t evaluated whether caffeine affects the results of tests of vestibular function was published by Felipe et al in 2005. The study evaluated whether the results from calorics, as well as tests of oculomotor function, were different depending on w hether th eir subjects consumed th e ir norm al amount of daily caffeine. The subjects in their study con sisted of 19 females who reported experiencing vestibular symptoms. The study found that the results from these two tests were similar across both test sessions. In a related study, Enriquez et al (2009) found that caffeine did not have any effect on postural stability in healthy young controls when they were asked to stand on a plat form as still as possible for 30 sec with their eyes open and 30 sec with their eyes closed. The purpose of the present study is to investigate if and how caffeine influences the outcome of two common clinical tests of vestibular func tion in healthy young adults, as patients are often asked to abstain from caffeinated foods and beverages 24 hr prior to vestibular testing (e.g., online instructions provided by the following practices: BayCare Clinic, Capital Region Otolaryngology Head and Neck Group, ENTCare).
P
METHODS hirty healthy controls (mean = 23.28, SD = 1.95 yr; 21 females) were tested during two separate sessions. During the caffeine (C) session, subjects
T
SBS
were asked to refrain from caffeine the morning of the test, and when they arrived for the test, they were asked to consume 16 oz. of Starbucks Breakfast Blend coffee, which contained ~300 mg of caffeine immedi ately prior to undergoing caloric and cVEMP testing (McCusker et al, 2003; www.starbucks.com). Subjectswere allowed to add cream and sugar to their coffee if they did not prefer to drink it black. During the no caffeine (NC) session, subjects were asked to refrain from caffeine for 24 hr prior to testing. The two sessions were counterbalanced. Subjects were free of any significant medical history, as well as free of any history of vestibular or balance disorders, and were compensated for their par ticipation. Approval for the study was obtained from the University at Buffalo institutional review board, and writ ten consent was obtained from each of the subjects. When subjects first arrived for the NC session, they were asked to complete a caffeine withdrawal question naire, which allowed us to assess the types of symptoms they were experiencing as a result of not drinking their normal amount of caffeine. The questionnaire assessed a number of symptoms, including headache, tiredness, difficulty concentrating, nausea/vomiting, and anxiety, and was adapted from a study that evaluated the type of caffeine withdrawal symptoms that were experienced by individuals 48 hr after removing caffeine from their diet (Ozsungur et al, 2009). The subjects were asked to rate the symptoms they were experiencing on an 11-point scale; a score of 0 indicated that they were not experienc ing the symptom, while a score of 10 indicated that they would rate that symptom as severe. Subjects were also required to complete a caffeine diary, which allowed eval uation of the average amount of caffeine each of the subjects consumed per week. Subjects were required to include the number and type of beverages, including brand (e.g., Starbucks, Dunkin Donuts, Pepsi, Coke, Red Bull), as well as the total number of ounces they consumed of each beverage each day. Milligrams per week of caffeine consumption was calculated for each subject through the use of the caffeine “dose by product” data reported on www. energyfiend.com, wilstar.com, and www.mayoclinic.com. Initially, subjects underwent tympanometry via the Audio Traveler AA222, which assessed the integrity of the middle ear. This was to ensure that each of the sub jects had normal middle ear function, as even a small con ductive hearing loss (resulting in an air-bone gap of 10 dB or more) can result in a reduced or absent cVEMP to airconducted stimuli (Halmagyi et al, 1994). In addition, tympanometry was performed to assure that the subject did not have a perforated eardrum prior to performing caloric testing. During caloric testing, subjects’ ears were irrigated with warm (44°C) and cool (30°C) w ater via the Variotherm Plus Caloric Irrigator (Atmos, Medizin Tecknik, Lenzkirch, Germany). Eye movement was monitored via in frared goggles. The I-P ortal VNG software
Influence of Caffeine on Calorics and cVEMPs/McNerney et al
(NeuroKinetics, Pittsburgh, PA) was used to collect and analyze the subjects’ eye movements. The percentages of unilateral weakness, directional preponderance, and asymmetry (degrees/second) were recorded and analyzed for each subject, as well as the PSV (degrees/second) for each of the four irrigations (i.e., right warm, left warm, right cool, and left cool). Unilateral weakness values of .007 (incorporating a Bonferroni correction), the only result that approached significance was the left warm irrigation. The Cohen’s d value for this paired comparison was equal to 0.48, which is considered to be a small to medium effect size. The raw data revealed that the PSV was slower during the C session than the NC session, which was also true of the remaining caloric irrigations (see Table 2). Despite the mean PSVs being larger in the NC session than in the C session, the mean directional preponderance and asymmetry results were larger in the C session than the NC session, while the unilateral weakness results were almost identical (i.e., 10.45 [C] vs. 10.47 [NC]) (see Table 2). When evaluating the individual data, there were three subjects who displayed abnormal results on calorics. All three displayed a unilateral weakness (UW). Two sub jects displayed a weakness that persisted throughout both sessions. The results from the two subjects were as follows: S20 (who was classified as a high caffeine
user) during the C session, UW = 43.46%, and during the NC session, UW = 38.50%, and S25 (who was clas sified a moderate user of caffeine) during the C session, UW = 35.95%, and during the NC session, UW = 35.27%. The third subject (S26, who was classified as a low caffeine user) only displayed a weakness dur ing the caffeine session, and it was ju st outside of normal limits (UW = 25.72%). During the NC session, the results were within normal limits for this subject (UW = 22.40%).
cVEMP Figure 1 displays an example of the cVEMP from an individual subject during the C and NC sessions. Table 3 fists the P I and N1 latencies as well as the P1N1 ampli tudes, for the right and left sides, during the C and NC sessions. The results reveal that there were minimal differ ences in mean cVEMP latency and amplitude between the C and the NC sessions. In addition, paired t-tests did not reveal any significant differences between the two ses sions. Clinically, an asymmetry ratio is often calculated, which allows one to determine if there is any difference' in the function of the right and left sides of a patient. Asymmetiy ratios up to 35% are considered to be within normal limits (Welgampola and Colebatch, 2001, 2005). Table 4 fists the asymmetry ratios for all of the subjects for both the C and NC sessions. The present study found mean asymmetry ratios of 15.06% (SD = 14.22) for the C session and 19.33% (SD = 13.33) for the NC session. In addition, statistical analysis of the asymmetry ratios via paired t-tests did not reveal any statistical difference between the C and NC sessions (t(29) = —1.712, p = 0.098, Mainr rence 4.05, S I!difi'ercRce 12.95, d .31). However, when evaluating the individual data, it was found that five subjects had values outside of normal limits. One sub ject (S30) fell outside of normal limits throughout both ses sions (C session: 54.40%; NC session: 49.99%), while the remaining four abnormal asymmetry ratios were as fol lows: S14: 52.81% (C), 25.99% (NC); S16: 37.43% (C), 19.27% (NC); S20: 33.69% (C), 51.53% (NC); S28: 3.41% (C), 44.94% (NC).
DISCUSSION any clinicians ask patients to refrain from con suming caffeine prior to undergoing tests of vestib ular function (ENTCare, Capital Region Otolaryngology Head and Neck Group, BayCare Clinic). The present study revealed that caffeine did not clinically or statisti cally (using a conservative Bonferroni approach) affect the test results between the C and NC sessions. There fore, the authors of the present study conclude that mod erate amounts of caffeine do not substantially affect the clinical interpretation of caloric or cVEMP testing in healthy young adults.
M
SG 4 J
Influence of Caffeine on Calorics and cVEMPs/McNerney et al
Table 2. Mean Peak Slow W ave Velocity (PSV) for Each of the Four Caloric Irrigations, as Well as the Percentage of Unilateral Weakness (UW), Directional Preponderance (DP), and Asymmetry Left cool
Right cool
Left warm
Right warm
DP (%)
UW (%)
Asymmetry (degrees/second)
10.45 8.58 -22.84 20.71 20.97 -20.44 Caffeine 10.47 7.15 -23.92 23.05 21.79 -20.94 No caffeine 0.02 1.43 1.08 2.34 0.82 0.50 Difference .912 .408 .361 .016 .184 .674 P value N o te : The difference row indicates the absolute mean difference between the C and NC sessions. Paired r-test results are row. Following a Bonferroni correction, p s 0.007 was considered to be significant.
Previous studies in our laboratory have also found that caffeine has minimal effects on the sensory orga nization test (SOT) (McNerney et al, forthcoming). Although the study found a statistically significant dif ference between the C and NC sessions for condition 5 (C5) and the composite score, the clinical interpretation of the results was not affected. In addition, the study also evaluated the change in C5 and composite score as a function of average daily caffeine consumption. The results showed that there was a greater change between the sessions for individuals who consumed lower amounts of caffeine as compared to those who consumed higher amounts of caffeine. This would strengthen the argument that (at least for healthy young controls) if individuals are accustomed to drinking moderate amounts of caffeine, asking them to refrain from doing so before undergoing tests of vestibular function is unnecessary. In the present study it is interesting to note that two of the individuals who did not report any history of vertigo or imbalance were found to have a UW, which persisted throughout both sessions. It is important to keep in mind that the variance for calorics is large and the caloric response is dependent on several factors, such as atten tion, the effectiveness of caloric stimulation, and size of the ear canal (Proctor and Glackin, 1985). Therefore, it is not all that improbable that two out of 30 subjects were outside of normal limits. In addition, when we evaluated the cVEMP asymme try ratios, we found that three subjects in the C session NC session
4.18 4.01 0.17 .738 listed in the bottom
revealed an abnormal asymmetry ratio (>35%) and three subjects in the NC session revealed an abnormal asymmetry ratio. Only one of the subjects had an abnor mal asymmetry ratio that persisted across session: (C: 54.4%; NC: 49.99%). Again, this subject did not report any vestibular or balance impairments prior to enrolling in the study. With the other three subjects, because the asymmetry ratios resolved in the opposite session, we can conclude with some certainty that the results are most likely due to the variability th at occurs with cVEMP amplitude. Published research studies have shown that P1N1 amplitude is highly variable and can be influenced by factors such as muscle tension, muscle fatigue, stimulus level, and how the subject is asked to contract the SCM muscle (Akin and Murnane, 2001; Ochi et al, 2001; Isaacson et al, 2006). In the present study, all of the subjects were required to monitor the amount of muscle tension they were producing during the cVEMP recordings, and were required to produce similar amounts of tension during both sessions. How ever, even small differences in the placement of the EMG monitoring electrode may have contributed to the differences in cVEMP amplitude seen across ses sion in these individuals. None of the subjects displayed abnormal results on both caloric and cVEMP tests, with the exception of one subject who displayed an abnormal cVEMP during the NC session; however, this asymmetry ratio resolved during the C session. This finding would be expected, C session
Figure 1. cVEMP recordings from the C and NC sessions from an individual subject. The NC session resulted in an asymmetry ratio of 9.86%, while the C session resulted in an asymmetry ratio of 23.85%.
S 65
J o u r n a l o f th e A m erican A cad em y o f A udiology/Volum e 25, N um ber 3, 2014
Table 3. Mean ± Standard Deviation for the Latency of P1 and N1 as Well as for the P1N1 Amplitude of the VEMP Latency (msec)
Caffeine No caffeine P value
Amplitude (pV)
P1_R
N1_R
P1_L
N1_L
P1N1_R
P1N1_L
13.46 ± 1.1 13.62 ± 1.2 .565
20.68 ± 1.8 20.44 ± 1.6 .471
13.94 ± 1.2 14.13 ± 1.2 .342
20.97 ± 1.8 20.95 ± 1.4 .488
187.63 ± 115.6 206.54 ± 123.4 .963
194.76 ± 148.6 189.72 ± 122.9 .745
Note: R - responses were recorded from the right SCM; L = responses were recorded from the left SCM.
given that the two tests are assessing two separate end organs of the vestibular system. Felipe et al (2005) is one of the only published studies that evaluated whether caffeine affects the results of calorics and oculomotor function. Similar to the present study, they also reported that there were minimal differ ences between the C and NC sessions for a group of 19 women ages 21 to 76 yr (Felipe et al, 2005). With regard to the oculomotor testing, they found that all of the results were normal, regardless of whether subjects were tested with or without caffeine. With regard to caloric testing, they reported a similar number of abnormalities in the C and NC sessions. Restated, if the subjects dis played an abnormal caloric result during the NC session, then they were found to display an abnormal caloric result during the C session. Felipe et al (2005) also reported that individuals expe rienced more anxiety as well as headache from cessation of caffeine. One limitation of the study by Felipe et al (2005) was that everyone in the study received the NC session first. Therefore, individuals could have naturally had more anxiety due to the fact that they were in a new situation and undergoing an unfamiliar test. The present study found that only four out of the 30 subjects indi cated that they were experiencing anxiety during the NC session. Three individuals rated their anxiety level as a one (mild), and one individual rated their anxiety level as a five (moderate). Three out of the four individ uals were in the C session first (particularly the one who rated her anxiety as a five), which means that when they underwent the second (NC) session, they already knew what to expect during each of the tests. Therefore, their anxiety is most likely related to the cessation of caf feine rather than an unfamiliar situation. Headache, on the other hand, was one of the more commonly reported symptoms on our symptom questionnaire. Eleven out of the 30 subjects reported experiencing this symptom as a result of abstaining from caffeine. Subjects rated this symptom from one to four. In addition to anxiety and headache, Felipe et al (2005) also reported that individ uals experienced more nausea and vomiting, as well as more intense vertigo, during the NC session. The present study found that four out of the 30 subjects reported nau sea as a caffeine withdrawal symptom. Two of the sub jects reported a rating of 1, one subject reported a rating of 2, and the final subject reported a rating of 3. Interest ingly, these were the same four subjects who reported
266
that they felt anxious as a result of cessation of caffeine. The conclusions of the present study regarding the caf feine withdrawal questionnaire would have been stron ger if we would have collected another questionnaire on the day that the subjects were undergoing the C session in order to account for symptoms that they may normally experience on a day-to-day basis. CONCLUSIONS
T
he present study evaluated whether the cessation of caffeine is necessary in order to undergo two
Table 4. VEMP Asymmetry Ratios (ARs) for Each of the Subjects during the Caffeine (C) and No-Caffeine (NC) Sessions S2 S3 S4 35 S6 S7 S8 S9 S10 S11 S12 S13 S14 S15 S16 S17 S18 S19 S20 S21 S22 S23 S24 S25 S26 S27 S28 S29 S30 S31
AR_C
AR_NC
0.62 2.19 29.27 12.32 11.20 12.53 5.70 13.14 1.01 10.50 9.02 6.72 52.81* 12.85 37.43* 10.74 11.83 21.37 33.69 0.80 0.94 23.85 21.85 7.23 8.86 24.74 3.41 6.05 54.40* 11.51
8.31 9.08 21.66 13.73 11.05 35.03 7.57 11.02 17.96 13.38 21.24 8.01 25.99 10.51 19.27 1.82 11.47 32.20 51.53* 9.90 1.61 9.86 30.70 19.55 27.55 30.02 44.94* 10.05 49.99* 14.99
*A value that is outside of normal limits.
Influence of Caffeine on Calorics and cVEMPs/McNerney et al
commonly used tests of vestibular function in young healthy adults. The results revealed that the ingestion of a moderate amount of caffeine did not substantially influence the clinical interpretation of either test. This would support the argument that young healthy adults do not need to refrain from drinking caffeine prior to caloric or cVEMP testing. In addition, the results from the caffeine withdrawal questionnaire indicate that the abrupt cessation of caffeine consumption may result in withdrawal symptoms such as headache, nausea, tired ness, and anxiety. In most cases, individuals who are experiencing vestibular symptoms are uncomfortable to begin with. The results from the present study indi cate that we may be causing greater discomfort to our patients than necessary if there are no effects of caf feine on the results from caloric and cVEMP tests. Future studies should focus on whether these same results occur in older adults and in individuals with com promised vestibular function and whether the results are affected by larger amounts of caffeine.
Enriquez A, Sklaar J, Viirre E, Chase B. (2009) Effects of caffeine on postural stability. Int Tinnitus J 15(2):161-163. Felipe L, Simoes LC, Gonjalves DU, Mancini PC. (2005) Evalua tion of the caffeine effect in the vestibular test. Braz J Otorhinolaryngol 71(6):758-762. Ferber-Viart C, Dubreuil C, Duclaux R. (1999) Vestibular evoked myogenic potentials in hum ans: a review. Acta Otolaryngol 119(1):6-15. Halmagyi GM, Colebatch JG, Curthoys IS. (1994) New tests of ves tibular function. Baillieres Clin Neurol 3(3):485—500. Isaacson B, Murphy E, Cohen H. (2006) Does the method of ster nocleidomastoid muscle activation affect the vestibular evoked myogenic potential response? J Vestib Res 16(4-5):187-191. McCusker RR, Goldberger BA, Cone EJ. (2003) Caffeine content of specialty coffees. J Anal Toxicol 27(7):520-522. McNemey KM, Coad ML, Burkard RF. (Forthcoming) The influence of caffeine on the sensory organization test (SOT). J Am Acad Audiol. Nolan S, Heinzer T. (2011) Essentials o f Statistics for the Behav ioral Sciences. 1st ed. New York: Worth Publishers.
REFERENCES
Ochi K, Ohashi T, Nishino H. (2001) Variance of vestibular-evoked myogenic potentials. Laryngoscope lll(3):522-527.
Akin FW, Murnane OD. (2001) Vestibular evoked myogenic poten tials: preliminary report. J Am Acad Audiol 12(9):445-452.
Ozsungur S, Brenner D, El-Sohemy A. (2009) Fourteen welldescribed caffeine w ithdraw al symptoms factor into three clus ters. Psychopharmacology (Berl) 201(4):541-548.
Akin FW, M urnane OD. (2008) Vestibular evoked myogenic potentials. In: Jacobson GP, Shepard NT, eds. Balance Function Assessment and Management. San Diego: Plural Publishing.
Proctor L, Glackin R. (1985) Factors contributing to variability of caloric test scores. Acta Otolaryngol 100(3-4):161-171.
Barin K. (2008) Background and technique of caloric testing. In: Jacobson GP, Shepard NT, eds. Balance Function Assessment and Management. San Diego: Plural Publishing.
Welgampola MS, Colebatch JG. (2001) Vestibulocollic reflexes: normal values and the effect of age. Clin Neurophysiol 112(11): 1971-1979.
Cohen J. (1988) Statistical Power Analysis for the Behavioral Sciences. 2nd ed. Hillsdale, NJ: Lawrence Erlbaum.
Welgampola MSC, Colebatch JG. (2005) Characteristics and clin ical applications of vestibular-evoked myogenic potentials. Neurol ogy 64(10):1682-1688.
Colebatch JG, Halmagyi GM, Skuse NF. (1994) Myogenic poten tials generated by a click-evoked vestibulocollic reflex. J Neurol Neurosurg Psychiatry 57(2):190-197.
Zhou G, Cox LC. (2004) Vestibular evoked myogenic potentials: history and overview. A m J Audiol 13(2): 135-143.
2G7
Copyright of Journal of the American Academy of Audiology is the property of American Academy of Audiology and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.
