SCIENTIFIC INVESTIGATIONS

Transcription

1 SCIENTIFIC INVESTIGATIONS Sedating Medications and Undiagnosed Obstructive Sleep Apnea: Physician Determinants and Patient Consequences Brandon Lu, M.D. 1 ; Rohit Budhiraja, M.D. 2 ; and Sairam Parthasarathy, M.D. 2 1 Loyola University Medical Center, Maywood, IL; 2 Section of Pulmonary and Critical Care Medicine, Southern Arizona VA Health Care System and the Department of Medicine, Sleep and Arizona Respiratory Centers, The University of Arizona, Tucson, AZ Background: Sedative medications may be inadvertently prescribed to patients with undiagnosed obstructive sleep apnea (OSA) and may worsen daytime sleepiness. Study Objectives: To determine whether patients with undiagnosed OSA were prescribed sedative medications and whether such prescriptions increased the risk for traffic accidents. A secondary objective was to determine physician characteristics associated with such prescription practices. Design: Retrospective chart review. Telephone interviews of patients and physicians. Intervention: None. Patients: One hundred fifty-one consecutive patients at a sleep laboratory. Results: Forty-one of 137 (30%) patients with undiagnosed OSA had received prescriptions for sedating medications. Regression analysis identified self-report of sleepiness while driving (p=.05) and prescription for risperidone as independent risk factors for motor vehicle accidents (p=.005), while prescription of any sedative (excluding risperidone) tended to be associated with accidents (p=.10). In patients with severe OSA, prescription of sedating medications was associated with a greater risk for motor vehicle accidents than those without such prescriptions (relative risk=2.6; p=.04). In patients with prescription for sedating medications (n=41), the apnea-hypopnea index was directly proportional to the risk for motor vehicle accidents (r 2 =0.26; p=.001) suggesting a dose effect of severity of sleep-disordered breathing on risk for accidents. Physicians who did not usually treat patients with sleep disorders were more likely to prescribe sedatives to patients with undiagnosed OSA than were physicians with such expertise: neurologist, pulmonologist, or psychiatrist (52% vs 10%; relative risk=5.2; p=.02) Conclusion: Prescription of sedating medications may increase the risk of road accidents in patients with undiagnosed severe OSA, and such prescription practices are less likely to occur in physicians with expertise in sleep medicine. Keywords: Sleep apnea, sedatives, specialist, automobile, traffic accidents, sleep Citation: Lu B; Budhiraja R; Parthasarathy S. Sedating medications and undiagnosed obstructive sleep apnea: physician determinants and patient consequences. J Clin Sleep Med 2005;1(4): Disclosure Statement Drs. Parthasarathy, Lu, and Budhiraja have indicated no financial conflicts of interest. Submitted for publication May 17, 2005 Accepted for publication May 18, 2005 Address correspondence to: Sairam Parthasarathy, M.D., Section of Pulmonary and Critical Care Medicine, Southern Arizona Veterans Administration Health Care System, 3601 South 6th Avenue, Tucson, AZ 85723; Tel: ; Fax: ; spartha@arc.arizona.edu In a survey of 1,154 adults in the United States, nearly 8% of those with difficulties sleeping were prescribed a sedative. 1 In the same survey, 61% of adults in the United States reported that their physician did not inquire about sleep-related symptoms, and other studies have reported that physicians do not routinely screen their patients for symptoms of obstructive sleep apnea (OSA). 2 However, nearly 10% of those adults surveyed reported having pauses in breathing, a symptom of OSA. 1 Therefore, one can expect patients with undiagnosed OSA to receive a prescription for sedatives, but the prevalence of this practice is not known. It is important to recognize that sedatives may be administered to patients with undiagnosed OSA because studies have shown that sedatives may have deleterious effects on respiration during sleep and daytime alertness. Benzodiazepines may depress the arousal response to hypoxia and hypercapnia during sleep 3 and reduce genioglossal muscle tone, 4,5 with consequent worsening of OSA measured as apnea-hypopnea index (AHI), duration of apneas, or severity of oxygen desaturation. 6-8 Additionally, sedating medications may independent of effects on respiration have adverse effects on daytime cognition and alertness, which, in turn, may lead to a greater risk for motor vehicle accidents. We set out to determine whether sedatives or medications with significant sedating side effects were being prescribed to patients Commentary Follows on Pages with OSA and, if so, whether such prescription practices were associated with worsening apnea severity and daytime sleepiness manifesting as an increased risk for motor vehicle accidents. A secondary objective of the study was to determine physician characteristics that were associated with such prescription practices level of training, primary specialty, and familiarity with diagnosing OSA. Identification of such characteristics that predispose a physician toward prescription of sedating medications in patients with undiagnosed OSA would identify the at-risk patient or phy- 367

2 B Lu, R Budhiraja, and S Parthasarathy Table 1 Sedative Medications and Medications with Sedative Effects Medication Patients, no. (% of prescription) Sedatives Zolpidem 5 (9.6) Benzodiazepine 17 (32.7) Diphenhydramine 8 (15.4) Medications with sedative effects Trazadone 5 (9.6) Risperidone 6 (11.5) Amitriptyline 6 (11.5) Nefazodone 2 (3.8) Mirtazapine 2 (3.8) Olanzapine 1 (1.9) sician groups for future preventative measures. Patient Selection METHODS One hundred and fifty-one consecutive patients who were diagnosed with OSA, defined as having an AHI greater than 5 per hour, were identified from a list of polysomnograms performed between March and November of the year 2000 in the Edward Hines Jr. VA Hospital. The Institutional Review Board had approved the study and granted a waiver of documentation of the informed consent. All patients and physicians were advised as to the research purposes of the telephone interview and were asked if they were willing to participate in such a study. Participants were excluded from the study if they refused participation during the telephone interview or if their medical and pharmacy records were not accessible (n=14). Table 2 Sedating Medications and Risk for Motor Vehicle Accidents Level of obstructive sleep apnea Mild Moderate Severe (n=47) (n=36) (n=54) (Proportion of patients as %) Patients receiving sedatives Suffered MVA 2 (16) 4 (28) 8 (53) No accidents 10 (84) 10 (72) 7 (47) Patients NOT receiving sedatives Suffered MVA 5 (14%) 6 (27%) 8 (20%) No accidents 30 (86) 16 (73) 31 (80) Apnea-hypopnea 12 ± 5 30 ± 5 66 ± 18 index* Minimum oxygen 85 ± 6 79 ± ± 13 saturation* MVA refers to Motor vehicle accident *mean ± SD Statistical Analysis Age, body-mass index, AHI, minimum SpO 2, and number of road traffic accidents for the group of patients who received sedating medications was compared with patients who did not receive such prescriptions using an unpaired Student t test. The differences in rates of motor vehicle accidents between patients with and without sedating medications were compared by χ 2 or Fisher exact tests, where appropriate. Simple and multiple logistic regression analysis were performed using STATISTICA software (StatSoft, Tulsa, OK). Data Collection The telephone questionnaire was aimed at determining the following: (1) use of prescription, over-the-counter, or herbal medications as sleep aids; (2) use of alcohol as a sedative; and (3) motor vehicle accidents attributable to sleepiness. The patient s chart was reviewed to verify and supplement the information obtained during the interview. Patient s age, body mass index, AHI, and minimum oxygen saturation (SpO 2 ) measured by pulse oximetry during the polysomnogram were obtained from the medical records. Patient medication lists were obtained from computerized records at the hospital pharmacy. Of medications that had sleepiness or drowsiness listed as side effects, only those that were refilled prior to the diagnostic polysomnogram, or were prescribed after the polysomnogram but before treatment of OSA was initiated, were considered for analysis (Table 1). For each patient, physicians who prescribed the sedative medication, or the primary care physicians caring for the patient, were contacted for a telephone-based interview of prescription practices. A total of 76 physicians responsible for such prescription practices were identified and agreed to participate. However, only 50 physicians could be reached or were willing to participate in the interview. Questions (Appendix) were aimed at determining their primary specialty, how often they prescribed sedating medications to patients with difficulties sleeping, awareness of contraindications to the use of sedating medications, how often they screened patients for OSA, and whether sleep disorders were part of their medical school curriculum. 368 RESULTS Of the 137 patients (age 59 ± 13 years) who were included in the analysis, the majority were men (n=132) and the body mass index was 35.1 ± 7.5 kg/m 2. The AHI and lowest SpO 2 during polysomnography were 39 ± 26 events/hour and 79% ± 12%, respectively. Sedating Medications Forty-one of 137 patients (30%) were prescribed sedating medications. Twenty patients (15%) were consuming alcohol, 16 patients (12%) were using over the counter medications, and 7 patients (5%) were ingesting herbal remedies to help them sleep. Patients receiving sedating medications were younger than patients who did not receive sedating medications: 53 ± 8 versus 61 ± 2 years (p=.02; odds ratio [OR] 1.05, 95% confidence interval [CI] ). Women were more likely than men to receive sedating medications. Four of 5 women (80%) received a prescription for a sedative medication, as opposed to 37 of 132 (28%) men (p=.039; relative risk [RR], 2.89). Body mass index and severity of OSA measured by the AHI or lowest Spo 2 were not different for patients with and without prescriptions for sedating medications (p=.6). Automobile Accidents Thirty-four percent of patients with prescriptions for sedating medications and 20% of patients without such prescriptions were involved in road traffic accidents (p=.21). We performed a

3 Sedatives, Accidents, and Sleep Expertise Table 3 Logistic Regression Analysis β SEM r 2 p Sedative prescription (including risperidone) Sedative prescription (excluding risperidone) Risperidone * Self-reported sleepy driving Sedative prescription (excluding risperidone) and risperidone Sedative Risperidone Sedative prescription (excluding risperidone), risperidone, sleepy driving Sedative Risperidone Sleepy driving subgroup analysis based on severity of OSA. Patients were divided into 3 groups based on AHI: mild (5-20 per hour), moderate (21-40), and severe (> 40)(Table 2). In patients with mild or moderate OSA, prescriptions for sedating medications did not confer any greater risk for suffering an automobile accident (p >.4). In patients with severe OSA, however, prescriptions for sedating medications were associated with an increased risk for automobile accidents (53% vs 20%, RR=2.6; p=.04)(figure 1). Moreover, in patients with prescriptions for sedating medications, a doseeffect of severity of OSA on proportion of patients suffering motor vehicle accidents was suggested; the proportion of patients suffering car accidents was 16%, 28%, and 53% for patients with mild, moderate, and severe OSA, respectively (Figure 1; p=.07; χ 2 ). In patients without prescriptions for sedating medications, there was no such dose effect: 14%, 27%, and 20% (Figure 1; p=.8). Similarly, in patients who consumed any form of sedative, prescription or nonprescription (including alcohol, over-thecounter, and herbal remedies), there tended to be an incremental proportion of patients suffering motor vehicle accidents: 13%, 38%, and 45% for mild, moderate, and severe OSA, respectively (χ 2 ; p=.08). In order to adjust for confounding variables that could influence the occurrence of traffic accidents, we performed univariate logistic regression analysis with occurrence of motor vehicle accidents as the dependent outcome variable and the following independent determining variables: ingestion of individual sedating medications, age, sex, and the use of herbal remedies, alcohol, and over-the-counter medications as sleep aids (n=137). Based on such univariate comparisons, we identified the following significant variables prescription of sedating medication (β=0.16; r 2 =0.03; p=.057), self-reported sleepiness while driving (β=0.16; r 2 =0.02; p=.057) and ingestion of risperidone (β=0.22; r 2 =0.05; p=.009). Age, sex, minimum SpO 2, and AHI were not associated with occurrence of traffic accidents (all p values >.20). Also, we performed analysis with the number of sedating medications as an independent variable. However, such analysis yielded no relationship between the number of sedating agents (ranging from 0 to 4 per patient) and motor vehicle accidents (r 2 = 0; p=.9; univariate logistic regression). Subsequent to identifying these significant independent variables, we built a multiple logistic regression model with motor vehicle accidents as the dependent variable (n=137)(table Table 4 Characteristics of Physicians Parameter Prescribed No sedative p value sedatives prescribed (n=23) (n=27) Years since graduation 10.4 ± ± from medical school Physician perceptions, percentage of time that he or she Prescribes sedatives 30.9 ± ± Screens patients for OSA 29.3 ± ± Medical school curriculum included information on sleep disorders, % Yes Knowledgeable about contraindication to the use of sedative medication, % Yes OSA refers to obstructive sleep apnea 3). Prescription for sedating medications, self-reported sleepiness while driving, and prescription for risperidone were independently associated with increased likelihood for motor vehicle accidents (Table 3). Physician Characteristics Of the 50 physicians who were surveyed regarding their prescribing patterns, 23 (46%) had prescribed sedating medications. There was no difference between the physicians who prescribed and those who had not prescribed sedating medications with respect to years since graduation from medical school, self-reported frequency of prescribing sedative medications, diligence in screening patients for OSA, knowledge of medication side effects, or presence or absence of sleep disorders in their medical school curriculum (Table 4; p>.25). Physicians who did not usually treat patients with sleep disorders were more likely to prescribe sedatives to patients with undiagnosed OSA than were physicians with such expertise-neurologist, pulmonologist, or psychiatrist (52% vs 10%, RR=5.2, p=.02)(figure 2). DISCUSSION To our knowledge, this is the first study to systematically examine the prevalence, and effect, of sedative prescription practices in patients with undiagnosed OSA. Some general observations can be made. First, the proportion of patients with undiagnosed OSA who are being prescribed sedating medications is significant 30%. Second, prescriptions for sedating medications may be associated with an increased risk for motor vehicle accidents in patients with severe OSA. Third, sedating medications, prescription or nonprescription, were not associated with worsening of apnea severity based upon SpO 2 or AHI. Fourth, physicians whose primary specialty did not include the practice sleep medicine were more likely to prescribe sedating medications to patients with undiagnosed OSA than were physicians belonging to specialties with sleep medicine expertise. A significant percentage of patients with undiagnosed OSA were prescribed a sedating medication. While only 8% of the general population is prescribed sedating medications, 1 we found that 30% of patients with OSA had such prescriptions. Patients with 369

4 B Lu, R Budhiraja, and S Parthasarathy Proportion of patients suffering accidents, % undiagnosed OSA may complain of difficulty sleeping at night, prompting physicians to prescribe such sedative medications. In our study, sedating medications during polysomnography did not influence apnea severity. During polysomnography, the AHI and lowest SpO 2 were not different between patients with and without prescriptions for sedating medications. Dolly and colleagues, 7 however, have reported higher AHI and lower SpO 2 in healthy subjects who received benzodiazepines. One reason for the difference between these 2 studies can be explained by the fact that some patients in our study did not receive the sedative until after their nocturnal polysomnogram; however, this constituted only 2% of all patients. Hence, other reasons, such as patients not ingesting the medications on the night of the polysomnogram, may explain such differences. Moreover, other investigators have not observed any worsening of sleep-disordered breathing when healthy subjects received sedating medications, such as oral hydromorphone. 8 These latter investigators have explained their findings by noting that preservation of the hypercapnic ventilatory response and pharyngeal patency in patients receiving the hydromorphone may have prevented the appearance of sleep-disordered breathing despite decrements in the hypoxic ventilatory response. 8 While most such were have been only a few studies in patients with OSA. 6,9 In a randomized cross-over study, patients with severe OSA experienced slight worsening of apnea-hypopnea duration and oxygen desaturation when administered triazolam, but such changes were mild and not clinically significant. 9,10 Alternatively, a cross-over study of zopiclone in patients with upper airway resistance syndrome revealed no worsening of sleepdisordered breathing. 11 In our study, the type of sedating medications prescribed may have contributed to the lack of worsening in measures of apnea severity (Table 1). Unlike events of sleep-disordered breathing, the effect of sedating medications on daytime vigilance and risk for motor vehicle accidents appears to be more uniform. In a recent study, commercial-vehicle drivers consuming narcotic analgesics or antihistamines were more likely to report motor vehicle accidents regardless of the presence or absence of OSA. 12 However, in this study, Howard and colleagues studied only commercial drivers. Furthermore, there was no interaction between the severity of * No sedatives Sedatives 0 Mild Moderate Severe Figure 1 The incidence of self-reported motor vehicle accidents due to sleepiness in patients with mild (apnea-hypopnea index 5-20 per hour), moderate (21-40 per hour) and severe (> 40 per hour) obstructive sleep apnea are shown. Patients with (black columns) and without (white columns) prescriptions for sedatives are shown. In patients with severe obstructive sleep apnea, the risk for traffic accidents is greater in those with prescriptions for sedative medications, as compared to those without such prescriptions (relative risk 2.6, p=.04). 370 Figure 2 Differences in proportion of sedative prescriptions issued by physicians whose primary specialty cared for patients with sleep disorders (white column) and other physicians (all other specialties, including internal medicine). Physicians who are trained in the specialties that are more likely to care for patients with sleep disorders were less likely to issue a prescription sedative (10%) to patients with undiagnosed obstructive sleep apnea than were other physicians (52%)(relative risk=5.2, p=.02; n=50). OSA and sedative ingestion reported, and not all patients suffered from OSA. In our study, the subjects are representative of patients usually referred to a sleep laboratory, and we found a dose effect for the effect of sedating medications on risk for motor vehicle accidents in patients with OSA (Figure 1). Patients with progressively worse OSA were more likely to have motor vehicle accidents when they had prescriptions for sedating medications. In contrast, when patients were not prescribed such sedating medications, there was no such dose effect. Although, such lack of relationship between severity of OSA and risk for traffic accidents has been observed by others, the combined effect of sedatives and OSA on traffic accidents has not been reported previously. 12,13 Such a dose effect persisted even when exposure to nonprescription medications such as over-the-counter medications, alcohol, and herbal remedies were considered. Conceivably, sedating medications may have blocked compensatory mechanisms operative in patients with sleep apnea and unmasked a preexisting dose effect of OSA on alertness. In the elderly, risperidone has been known to cause lack of coordination and falls. 14 Risperidone has a long half-life (20 hours) and has sedating actions that may be potentiated by concomitant ingestion of herbs and alcohol. The only other agents consumed by our patients with a similar half-life were mirtazapine (20-40 hours), olanzapine (21-40 hours), and clonazepam (19 hours). Interestingly, even after we adjusted for other variables in a multiple logistic regression model, Risperidone remained independently associated with motor vehicle accidents. The long half-life of the medication and the effects on coordination may be responsible for such an effect. Patient outcomes in certain disorders can differ based on physicians primary specialty. 15,16 However, such differences in patient outcomes based on physician expertise have not been described in sleep medicine. 17 Lately, such an area of research is receiving considerable attention from the American Academy of Sleep Medicine. 17 In our study, physicians belonging to primary specialties that customarily provide care to patients with sleep disorders pulmonary, neurology, and psychiatry were less likely

5 Sedatives, Accidents, and Sleep Expertise to prescribe sedating medications to patients with undiagnosed OSA. The prescribing of sedating medications may have resulted in poor patient outcomes, as documented by patients with severe OSA in our study experiencing a greater number of motor vehicle accidents than similar patients without prescriptions for sedating medications. Additionally, female patients with undiagnosed OSA appear to be at a greater risk of being prescribed sedating medications. However, the number of women in our study was small. Nevertheless, this finding may reflect the universally consistent finding of more insomnia complaints in women. 18 Future education efforts may need to target such physician and patient populations to prevent prescription of sedating medications to patients who are at risk for motor vehicle accidents. Interestingly, only a third of the physicians screened their patients for OSA (Table 4). Hence, factors other than the likelihood for undiagnosed OSA must have determined physicians prescription practices. Age and sex of patients may have played a role. Greater age of patients and consequent concern for falls may have deterred physicians from prescribing sedating medications to the aged. 19 Also, women are more likely to complain of insomnia than men, and this may explain the greater likelihood for them to receive sedating medications. 18,20 Critiques and Limitations There are several limitations to our study. We did not control for driving hours or the number of accidents per patient. However, the presence of a dose effect of AHI on the risk for accidents speaks against such a systematic bias. Also, we relied on physicians memory in answering questions regarding their prescription practices, and this may be prone to recall bias. However, review of physician notes does not always reveal the reasons for their decision making and, thus, scrutinizing the charts for such documentation may not have provided the necessary additional information. Some of the other limitations of our study are the small numbers, retrospective study design, and self-report of accidents instead of verification from department of motor vehicles records. However, while the total rates of accidents may have been underestimated when measuring self-report of accidents instead of accident records, we doubt that the results would have been systematically biased to favor one or the other comparison groups patients with and without access to sedating medication. Additionally, this study was performed in a veteran s population, and, hence, there is underrepresentation of women html#8 National Sleep Foundation: 2000 Omnibus Sleep in America Poll (Last accessed on May 19, 2005). 2. Rahaghi F, Basner RC. Delayed Diagnosis of obstructive sleep apnea: don't ask, don't tell. Sleep Breath 1999;3: Hedemark LL, Kronenberg RS. Flurazepam attenuates the arousal response to co 2 during sleep in normal subjects. Am Rev Respir Dis 1983;128: Bonora M, Shields GI, Knuth SL, Bartlett D Jr, St John WM. Selective depression by ethanol of upper airway respiratory motor activity in cats. Am Rev Respir Dis 1984;130: Leiter JC, Knuth SL, Bartlett D, Jr. The effect of sleep deprivation on activity of the genioglossus muscle. Am Rev Respir Dis 1985;132: Mendelson WB, Garnett D, Gillin JC. Flurazepam-induced sleep apnea syndrome in a patient with insomnia and mild sleep-related respiratory changes. J Nerv Ment Dis 1981;169: Dolly FR, Block AJ. Effect of flurazepam on sleep-disordered breathing and nocturnal oxygen desaturation in asymptomatic subjects. Am J Med 1982;73: Robinson RW, Zwillich CW, Bixler EO, Cadieux RJ, Kales A, White DP. Effects of oral narcotics on sleep-disordered breathing in healthy adults. Chest 1987;91: Berry RB, Kouchi K, Bower J, Prosise G, Light RW. Triazolam in patients with obstructive sleep apnea. Am J Respir Crit Care Med 1995;151: George CF, Boudreau AC, Smiley A. Comparison of simulated driving performance in narcolepsy and sleep apnea patients. Sleep 1996;19: Lofaso F, Goldenberg F, Thebault C, Janus C, Harf A. Effect of zopiclone on sleep, night-time ventilation, and daytime vigilance in upper airway resistance syndrome. Eur Respir J 1997;10: Howard ME, Desai AV, Grunstein RR, Hukins C, Armstrong JG, Joffe D, Swann P, Campbell DA, Pierce RJ. Sleepiness, sleep-disordered breathing, and accident risk factors in commercial vehicle drivers. Am J Respir Crit Care Med 2004;170: Barbe, Pericas J, Munoz A, Findley L, Anto JM, Agusti AG. Automobile accidents in patients with sleep apnea syndrome. An epidemiological and mechanistic study. Am J Respir Crit Care Med 1998;158: Katz IR, Rupnow M, Kozma C, Schneider L. Risperidone and falls in ambulatory nursing home residents with dementia and psychosis or agitation: secondary analysis of a double-blind, placebo-controlled trial. Am J Geriatr Psychiatry 2004;12: Jong P, Gong Y, Liu PP, Austin PC, Lee DS, Tu JV. Care and outcomes of patients newly hospitalized for heart failure in the community treated by cardiologists compared with other specialists. Circulation 2003;108: Nash IS, Nash DB, Fuster V. Do cardiologists do it better? J Am Coll Cardiol. 1997;29: Grant opportunities. AASM Bull. Winter ; Klink M, Quan SF, Lebowitz MD, Kaltenborn W. Risk factors for complaints of insomnia in a general adult population. Arch Intern Med 1992; e152: Stenbacka M, Jansson B, Leifman A, Romelsjo A. Association between use of sedatives or hypnotics, alcohol consumption, or other risk factors and a single injurious fall or multiple injurious falls: a longitudinal general population study. Alcohol 2002;28: Quintana-Gallego E, Carmona-Bernal C, Capote F, et al. Gender differences in obstructive sleep apnea syndrome: a clinical study of 1166 patients. Respir Med 2004;98: AKNOWLEDGEMENTS Supported by a grant from Richard M. Bane Trust and a Southern Arizona VA Healthcare and University of Arizona Research Award. REFERENCES 371 Physician Questionnaire APPENDIX 1. Level of training. If in training: PGY year? If practicing: year graduating medical school? 2. Specialty? 3. If a patient complains of inability to sleep, how often do you prescribe sedatives to such a patient? Never 25% 50% 75% Always 4. What medical conditions do you consider are contraindications for sedatives? 5. Do you screen all your patients for obstructive sleep apnea? Never 25% 50% 75% Always 6. Did your medical school curriculum include sleep disorders? Yes No don t know