Portal de Periódicos da CAPES

Symptoms of depression, prescription of benzodiazepines, and the risk of death in hemodialysis patients in Japan

2006; Elsevier BV; Volume: 70; Issue: 10 Linguagem: Inglês

10.1038/sj.ki.5001832

1523-1755

Shunichi Fukuhara, J. Green, Justin M. Albert, Hanako Mihara, R. L. Pisoni, Shin Yamazaki, T. Akiba, Tadao Akizawa, Yoshihide Asano, Akira Saitō, F. K. Port, Phillip J. Held, Kiyoshi Kurokawa,

Family Caregiving in Mental Illness

Many hemodialysis patients in Japan have symptoms of depression, but whether those patients are treated appropriately is unknown. As part of the Dialysis Outcomes and Practice Patterns Study, data on symptoms of depression, physician-diagnosed depression, prescribed medications, and death were collected prospectively in cohorts in Japan (n=1603) and 11 other countries (n=5872). Symptoms of depression were as prevalent in Japan as elsewhere, but in Japan a much smaller percentage of patients had physician-diagnosed depression: only 2% in Japan vs 17% elsewhere. Antidepressants were much less commonly prescribed in Japan: only 1% in Japan vs 17% elsewhere for patients with many and frequent symptoms of depression, and 16% in Japan vs 34% elsewhere for patients with physician-diagnosed depression. In Japan, symptoms of depression were associated with prescription of benzodiazepines (without antidepressants), and patients with physician-diagnosed depression were twice as likely to be given benzodiazepines: 32% in Japan vs 16% elsewhere. Benzodiazepine monotherapy was associated with death (relative risk 1.56, 95% confidence interval (CI), 1.25–1.94), even after adjustments for 13 likely confounders (relative risk 1.27, 95% CI, 1.01–1.59). Hemodialysis patients in Japan with symptoms of depression are given not antidepressants but benzodiazepines, a practice associated with higher mortality. Many hemodialysis patients in Japan have symptoms of depression, but whether those patients are treated appropriately is unknown. As part of the Dialysis Outcomes and Practice Patterns Study, data on symptoms of depression, physician-diagnosed depression, prescribed medications, and death were collected prospectively in cohorts in Japan (n=1603) and 11 other countries (n=5872). Symptoms of depression were as prevalent in Japan as elsewhere, but in Japan a much smaller percentage of patients had physician-diagnosed depression: only 2% in Japan vs 17% elsewhere. Antidepressants were much less commonly prescribed in Japan: only 1% in Japan vs 17% elsewhere for patients with many and frequent symptoms of depression, and 16% in Japan vs 34% elsewhere for patients with physician-diagnosed depression. In Japan, symptoms of depression were associated with prescription of benzodiazepines (without antidepressants), and patients with physician-diagnosed depression were twice as likely to be given benzodiazepines: 32% in Japan vs 16% elsewhere. Benzodiazepine monotherapy was associated with death (relative risk 1.56, 95% confidence interval (CI), 1.25–1.94), even after adjustments for 13 likely confounders (relative risk 1.27, 95% CI, 1.01–1.59). Hemodialysis patients in Japan with symptoms of depression are given not antidepressants but benzodiazepines, a practice associated with higher mortality. Depression in hemodialysis patients is underdiagnosed. In 12 countries participating in the Dialysis Outcomes and Practice Patterns Study (DOPPS), the number of patients with high scores on a standard questionnaire of symptoms of depression was more than three times the number in whom depression had been diagnosed by a physician.1.Lopes A.A. Albert J.M. Young E.W. et al.Screening for depression in hemodialysis patients: associations with diagnosis, treatment, and outcomes in the DOPPS.Kidney Int. 2004; 66: 2047-2053Abstract Full Text Full Text PDF PubMed Scopus (219) Google Scholar Depression is also undertreated: of the DOPPS patients with physician-diagnosed depression, only about one-third had been given a prescription for antidepressant medication, even though symptoms of depression were associated with withdrawal from treatment and with death.1.Lopes A.A. Albert J.M. Young E.W. et al.Screening for depression in hemodialysis patients: associations with diagnosis, treatment, and outcomes in the DOPPS.Kidney Int. 2004; 66: 2047-2053Abstract Full Text Full Text PDF PubMed Scopus (219) Google Scholar In Japan, underdiagnosis was much more common than in any of the 11 other DOPPS countries,1.Lopes A.A. Albert J.M. Young E.W. et al.Screening for depression in hemodialysis patients: associations with diagnosis, treatment, and outcomes in the DOPPS.Kidney Int. 2004; 66: 2047-2053Abstract Full Text Full Text PDF PubMed Scopus (219) Google Scholar but at the time of that study the data on medications given to DOPPS patients in Japan were not available. We obtained those data, and examined the prescription of antidepressants and benzodiazepines to hemodialysis patients in Japan. We also studied whether those prescriptions were associated with symptoms of depression, physician-diagnosed depression, hospitalization, and death. Symptoms of depression were common among patients in Japan: 40.0% had Center for Epidemiological Studies Depression Screening Index (CES-D) scores of 10 or greater, and 14.4% had scores of 15 or greater (Table 1). These did not differ markedly from the other DOPPS countries overall (43.9 and 19.5%, Table 1). Patients in Japan were no more or less likely than those in the other countries to have high CES-D scores (adjusted odds ratio, 0.97; 95% confidence interval (CI), 0.84–1.11).Table 1Symptoms of depression, physician-diagnosed depression, and medications in hemodialysis patients in Japan and in the 11 other DOPPS countriesaDOPPS II data (2002–2004), in prevalent cross-section of patients with information on physician-diagnosed depression, symptoms of depression, and medications (N=7475).MeasurePercent or mean±s.d. (n/N)JapanOther DOPPS countriesN=1603N=5872CES-D score9.1±5.2, median=89.3±5.9, median=8 0–418.0 (289/1603)22.9 (1346/5872) 5–941.9 (672/1603)33.0 (1938/5872) 10–1425.6 (410/1603)24.4 (1430/5872) 15–199.5 (153/1603)13.5 (795/5872) 20–244.2 (67/1603)5.0 (295/5872) 25–300.7 (12/1603)1.0 (59/5872)MHI-5 score65.3±21.9, median=65.0Not measuredPrevalence of CES-D score ≥1040.0 (642/1603)44.0 (2582/5872) MHI-5 score <6038.0 (564/1485)Not measured Physician-diagnosed depression1.9 (31/1603)17.4 (1020/5872) Physician-diagnosed depression in those with CES-D score ≥102.3 (15/642)25.2 (650/2582)Ratio of CES-D ≥10 to physician-diagnosed depression20.72.5Antidepressants Overall1.2 (19/1603)12.8 (750/5872) In those with CES-D score ≥10/<101.2 (8/642)/1.1 (11/961)17.2 (444/2582)/9.3 (306/3290) In those with/without physician-diagnosed depression16.1 (5/31)/0.9 (14/1572)33.7 (344/1020)/8.4 (406/4852)Benzodiazepines (without antidepressants) Overall19.2 (308/1603)14.8 (869/5872) In those with CES-D score ≥10/ 0.05), but there was slightly less disparity in institutions that had a department of psychiatry (psychiatry department present vs absent; 86.7 vs 95.0%, P<0.05). In Japan, very few patients were given antidepressants (only 19 of 1603 patients, 1.2%). This was also true of those with high CES-D scores (only eight of 642 patients, 1.2%). The comparable percentages for the other DOPPS countries were more than 10 times higher: 12.8 and 17.2% (Table 1, Figure 1b). Even in patients with physician-diagnosed depression, prescription of antidepressants was only half as common in Japan as in the other countries: 16.1 vs 33.7% (Table 1). Benzodiazepines were prescribed to about one-fifth of the patients in Japan and to about one-seventh of those in the other DOPPS countries (Table 1). In Japan, patients with physician-diagnosed depression were twice as likely as those in the other countries to be given benzodiazepines: 32.3 vs 15.7% (Table 1). Unlike in the other countries, in Japan patients who had high CES-D scores were almost twice as likely as those with lower scores to be given benzodiazepines: 26.6 vs 14.3% (Table 1). The association between CES-D scores and prescription of benzodiazepines was stronger in Japan than in the other countries (Figure 1c). Within Japan, patients who received benzodiazepines had lower five-item mental health index scores (indicating more symptoms of depression; adjusted odds ratio, 0.984, 95% CI, 0.980–0.988), but benzodiazepine prescription was not associated with albumin concentration (an indicator of end-stage renal disease severity; adjusted odds ratio, 0.906, 95% CI, 0.737–1.114). Benzodiazepine prescribing practice also varied among dialysis facilities in Japan. At nine of the 59 facilities studied, no more than 10% of the patients with high CES-D scores were given benzodiazepines, but at 11 facilities more than 40% of the patients with high CES-D scores were given these drugs (Figure 2). Japan had many facilities where patients with high CES-D scores were likely to be given benzodiazepines, but in the other countries such facilities were rare. Benzodiazepine prescribing practice was more uniform among dialysis facilities in the other countries than among facilities in Japan. CES-D scores were positively associated with the risk of hospitalization and of death (Figure 3). Having received a prescription for a benzodiazepine was not associated with the risk of hospitalization, but it was associated with the risk of death (Table 2). That association remained even after adjusting for depression, and after adjusting for demographic and socioeconomic characteristics, time on dialysis, comorbid conditions, albumin, eKt/V, and hemoglobin (Table 2). The risk of death was also greater at dialysis facilities that were above the median for the percentage of patients given benzodiazepines (hazard ratio, 1.36, 95% CI, 1.03–1.79).Table 2Associations between benzodiazepine prescription and the risks of death and of hospitalizationaJapan DOPPS I and II data (1999–2004, N=5041); models stratified by phase of DOPPS data collection, adjusted as described above, and controlled for effects of facility clustering.Adjustment levelRelative riskbRisks of hospitalization and of death in patients given prescriptions for benzodiazepines and not antidepressants, relative to the risk in patients given neither benzodiazepines nor antidepressants. (95% CI) P-valuesMortalityHospitalization1. Unadjusted1.56 (1.25–1.94)1.14 (1.01–1.28) 1.00–1.59)0.99 (0.87–1.12)socioeconomic variables, and comorbid conditions0.050.844. Same as no. 3 above, but also adjusted for albumin, eKt/V, and hemoglobin1.27 (1.01–1.59)0.99 (0.88–1.12)0.040.91CI, confidence interval; MHI-5, five-item mental health index.The numbers in italics are P-values.a Japan DOPPS I and II data (1999–2004, N=5041); models stratified by phase of DOPPS data collection, adjusted as described above, and controlled for effects of facility clustering.b Risks of hospitalization and of death in patients given prescriptions for benzodiazepines and not antidepressants, relative to the risk in patients given neither benzodiazepines nor antidepressants. Open table in a new tab CI, confidence interval; MHI-5, five-item mental health index. The numbers in italics are P-values. The discrepancy between symptoms of depression and physician-diagnosed depression is greater in Japan than in other countries (Table 1, Figure 1a).1.Lopes A.A. Albert J.M. Young E.W. et al.Screening for depression in hemodialysis patients: associations with diagnosis, treatment, and outcomes in the DOPPS.Kidney Int. 2004; 66: 2047-2053Abstract Full Text Full Text PDF PubMed Scopus (219) Google Scholar To interpret this finding, we note first that the social stigmatization in Japan of people given the diagnosis of depression is well known.2.Fukuhara S. Lopes A.A. Bragg-Gresham J.L. et al.Health-related quality of life among dialysis patients on three continents: the Dialysis Outcomes and Practice Patterns Study (DOPPS).Kidney Int. 2003; 64: 1903-1910Abstract Full Text Full Text PDF PubMed Scopus (210) Google Scholar, 3.Waza K. Graham A.V. Zyzanski S.J. Inoue K. Comparison of symptoms in Japanese and American depressed primary care patients.Fam Pract. 1999; 16: 528-533Crossref PubMed Scopus (49) Google Scholar, 4.Mino Y. Aoyama H. Froom J. Depressive disorders in Japanese primary care patients.Fam Pract. 1994; 11: 363-367Crossref PubMed Scopus (28) Google Scholar, 5.Froom J. Aoyama H. Hermoni D. et al.Depressive disorders in three primary care populations: United States, Israel, Japan.Fam Pract. 1995; 12: 274-278Crossref PubMed Scopus (19) Google Scholar, 6.Shima S. Satoh E. Somatoform disorders in the workplace in Japan.Int Rev Psychiatry. 2006; 18: 35-40Crossref PubMed Scopus (7) Google Scholar In addition, a reluctance to identify depression can be found among the general public: when Japanese and Australians were given a vignette of someone with clear signs and symptoms of clinical depression and then asked to describe that person's condition, the Japanese were only one-third as likely as the Australians to use the term ‘depression’.7.Jorm A.F. Nakane Y. Christensen H. et al.Public beliefs about treatment and outcome of mental disorders: a comparison of Australia and Japan.BMC Med. 2005; 3: 12Crossref PubMed Scopus (181) Google Scholar In Japan, symptoms of depression may even ‘be given positive social meanings’.8.Kirmayer L.J. Psychopharmacology in a globalizing world: the use of antidepressants in Japan.Transcultural Psychiatry. 2002; 39: 295-322Crossref Scopus (44) Google Scholar The Diagnostic and Statistical Manual is available in Japanese, but remaining to be studied in sufficient depth are the diagnostic criteria actually used by Japanese physicians when they encounter patients with symptoms of depression, how such practices differ between countries, and how they differ between physicians of different educational and clinical backgrounds. The relations of psychiatrists to their patients are changing,9.Slingsby B.T. Decision-making models in Japanese psychiatry: transitions from passive to active patterns.Soc Sci Med. 2004; 59: 83-91Crossref PubMed Scopus (24) Google Scholar and recognition and treatment of depression is improving among specialist providers of psychiatric services,10.Tajima O. Mental health care in Japan: recognition and treatment of depression and anxiety disorders.J Clin Psychiatry. 2001; 62 (Discussion 45–46): 39-44PubMed Google Scholar but nephrologists, even those who recognize clinical depression, might still not record the diagnosis. Symptoms of depression are often somatic,3.Waza K. Graham A.V. Zyzanski S.J. Inoue K. Comparison of symptoms in Japanese and American depressed primary care patients.Fam Pract. 1999; 16: 528-533Crossref PubMed Scopus (49) Google Scholar, 6.Shima S. Satoh E. Somatoform disorders in the workplace in Japan.Int Rev Psychiatry. 2006; 18: 35-40Crossref PubMed Scopus (7) Google Scholar, 11.Nakao M. Yano E. Reporting of somatic symptoms as a screening marker for detecting major depression in a population of Japanese white-collar workers.J Clin Epidemiol. 2003; 56: 1021-1026Abstract Full Text Full Text PDF PubMed Scopus (43) Google Scholar and because renal failure affects many systems, nephrologists’ task of distinguishing somatic symptoms of depression from ‘direct’ effects of renal failure is particularly complex. This might be alleviated if medical students and postgraduate trainees received better training in the diagnosis of mild and moderately severe psychiatric conditions. One might expect the presence of a psychiatrist to make the diagnosis of depression more likely, and the present data do support that idea: in patients with high CES-D scores the diagnosis of depression was more common at dialysis facilities in institutions with a department of psychiatry. Such facilities accounted for one-quarter of the facilities studied in Japan. Thus, the percentage of hemodialysis patients to whom psychiatric services were readily available was small. Of course, even if a psychiatrist is available and a patient is given a referral, the patient can still refuse consultation or treatment. The majority of people in Japan who experience a severe or moderate mental disorder do not seek medical treatment.12.Kawakami N. Takeshima T. Ono Y. et al.Twelve-month prevalence, severity, and treatment of common mental disorders in communities in Japan: preliminary finding from the World Mental Health Japan Survey 2002–2003.Psychiatry Clin Neurosci. 2005; 59: 441-452Crossref PubMed Scopus (248) Google Scholar One reason may be found in the 2006 report by Shima and Satoh,6.Shima S. Satoh E. Somatoform disorders in the workplace in Japan.Int Rev Psychiatry. 2006; 18: 35-40Crossref PubMed Scopus (7) Google Scholar who noted that the stigma mentioned above ‘continues to be associated with those suffering from mental disorders and for those who have to consult a psychiatrist’. In addition, among the general public in Japan psychiatrists are not commonly mentioned as people who would be helpful to someone with symptoms of depression: psychiatrists ranked fifth, behind counselors, close family, close friends, and social workers.7.Jorm A.F. Nakane Y. Christensen H. et al.Public beliefs about treatment and outcome of mental disorders: a comparison of Australia and Japan.BMC Med. 2005; 3: 12Crossref PubMed Scopus (181) Google Scholar In the other DOPPS countries, antidepressants were prescribed to one-third of the patients with physician-diagnosed depression, but in Japan the comparable fraction was about one-sixth (Table 1). That is, even after diagnosing depression, physicians in Japan were only half as likely to prescribe antidepressants as were physicians in other countries. Prescription of antidepressants is tied to socioeconomic and cultural forces,8.Kirmayer L.J. Psychopharmacology in a globalizing world: the use of antidepressants in Japan.Transcultural Psychiatry. 2002; 39: 295-322Crossref Scopus (44) Google Scholar,13.Kleinman A. Culture and depression.N Engl J Med. 2004; 351: 951-953Crossref PubMed Scopus (343) Google Scholar with the result that, in Japan, antidepressants are prescribed to patients with diagnosed depression at dosages that are below international standards.14.Furukawa T.A. Kitamura T. Takahashi K. Treatment received by depressed patients in Japan and its determinants: naturalistic observation from a multi-center collaborative follow-up study.J Affect Disord. 2000; 60: 173-179Abstract Full Text Full Text PDF PubMed Scopus (21) Google Scholar That being the case in psychiatric settings,14.Furukawa T.A. Kitamura T. Takahashi K. Treatment received by depressed patients in Japan and its determinants: naturalistic observation from a multi-center collaborative follow-up study.J Affect Disord. 2000; 60: 173-179Abstract Full Text Full Text PDF PubMed Scopus (21) Google Scholar physicians without psychiatric training may be even more reluctant to prescribe these drugs. Nephrologists might be unsure about the most appropriate doses and regimens, and about monitoring patients who receive antidepressants. The limited availability of psychiatric services (noted above) and reluctance to refer patients to outside institutions would make prescription of antidepressants even less likely. Nephrologists might also avoid prescribing antidepressants because of anticholinergic effects: dry mouth could interfere with the restriction of fluid intake. Increased fluid intake might be seen as more dangerous than suboptimal treatment of recognized depression, but why that would be true only in Japan is unclear. We also note that the hypothesis that increased use of antidepressants would improve outcomes remains to be tested. We could not directly test the hypothesis that benzodiazepines were substituted for antidepressants, but the results in Table 1 and in Figures 1b, c, and 2 show that patients who might be expected to benefit from antidepressants were given not antidepressants but benzodiazepines. The benzodaizepines prescribed without antidepressants to those patients may have been intended as treatment, however inappropriate, for depression. They might also have been prescribed to treat pruritus, insomnia,15.Novak M. Shapiro C.M. Mendelssohn D. Mucsi I. Diagnosis and management of insomnia in dialysis patients.Semin Dial. 2006; 19: 25-31Crossref PubMed Scopus (77) Google Scholar or anxiety. Even the possibility that some symptoms of depression reported by dialysis patients result from, or are aggravated by, benzodiazepine use should be considered.10.Tajima O. Mental health care in Japan: recognition and treatment of depression and anxiety disorders.J Clin Psychiatry. 2001; 62 (Discussion 45–46): 39-44PubMed Google Scholar,16.Nathan R.G. Robinson D. Cherek D.R. et al.Long-term benzodiazepine use and depression.Am J Psychiatry. 1985; 142: 144-145Crossref PubMed Scopus (5) Google Scholar In Japan, when members of the general public were asked which drugs would be helpful for someone with symptoms of depression, about 30% indicated ‘sleeping pills’ and about 40% indicated ‘tranquilizers’.7.Jorm A.F. Nakane Y. Christensen H. et al.Public beliefs about treatment and outcome of mental disorders: a comparison of Australia and Japan.BMC Med. 2005; 3: 12Crossref PubMed Scopus (181) Google Scholar This public opinion is consistent with Japanese medical practice. Specifically, benzodiazepines are prescribed not only to about half of patients with depression who are under psychiatric care,14.Furukawa T.A. Kitamura T. Takahashi K. Treatment received by depressed patients in Japan and its determinants: naturalistic observation from a multi-center collaborative follow-up study.J Affect Disord. 2000; 60: 173-179Abstract Full Text Full Text PDF PubMed Scopus (21) Google Scholar but also to patients with depressed mood who are seen in primary care.8.Kirmayer L.J. Psychopharmacology in a globalizing world: the use of antidepressants in Japan.Transcultural Psychiatry. 2002; 39: 295-322Crossref Scopus (44) Google Scholar,10.Tajima O. Mental health care in Japan: recognition and treatment of depression and anxiety disorders.J Clin Psychiatry. 2001; 62 (Discussion 45–46): 39-44PubMed Google Scholar Because benzodiazpines are not indicated in the treatment of depression per se,17.Mann J.J. Drug therapy: the medical management of depression.N Engl J Med. 2005; 353: 1819-1834Crossref PubMed Scopus (323) Google Scholar,18.Whooley M.A. Simon G.E. Primary care: managing depression in medical outpatients.N Engl J Med. 2000; 343: 1942-1950Crossref PubMed Scopus (334) Google Scholar one would expect to find only a few dialysis facilities where they were prescribed to large percentages of patients with high CES-D scores. That is exactly the pattern found in the countries other than Japan. In contrast, at many Japanese facilities benzodiazepines were prescribed to large percentages of such patients, which is consistent with the hypothesis that in Japan benzodiazepines are prescribed to treat symptoms of depression (Figure 2). The practice of prescribing benzodiazepines to patients with symptoms of depression varied more widely among dialysis facilities in Japan than in other countries. Whether the variation in prescribing practice within Japan is related to variation in case mix, physician training, or other characteristics of the facilities remains to be studied, as does the hypothesis that the nine Japanese facilities at the left end of the x axis in Figure 2 should serve as models of good prescribing practice. Having a high CES-D score is a risk factor for hospitalization1.Lopes A.A. Albert J.M. Young E.W. et al.Screening for depression in hemodialysis patients: associations with diagnosis, treatment, and outcomes in the DOPPS.Kidney Int. 2004; 66: 2047-2053Abstract Full Text Full Text PDF PubMed Scopus (219) Google Scholar,19.Hedayati S.S. Grambow S.C. Szczech L.A. et al.Physician-diagnosed depression as a correlate of hospitalizations in patients receiving long-term hemodialysis.Am J Kidney Dis. 2005; 46: 642-649Abstract Full Text Full Text PDF PubMed Scopus (96) Google Scholar and for death among hemodialysis patients worldwide.1.Lopes A.A. Albert J.M. Young E.W. et al.Screening for depression in hemodialysis patients: associations with diagnosis, treatment, and outcomes in the DOPPS.Kidney Int. 2004; 66: 2047-2053Abstract Full Text Full Text PDF PubMed Scopus (219) Google Scholar The present results show that symptoms of depression are important risk factors for those outcomes in Japanese patients in particular (Figure 3). Another new finding of the present study is that prescription of a benzodiazepine without prescription of an antidepressant is a risk factor for death. Benzodiazepines might have been prescribed as palliative care when patients were severely ill and already near death, but if such ‘confounding by indication’ did occur its effects were small: albumin concentration (an indicator of disease severity) was not associated with benzodiazepine prescription; facility-level analysis20.Wolfe R.A. Observational studies are just as effective as randomized clinical trials.Blood Purif. 2000; 18: 323-326Crossref PubMed Scopus (15) Google Scholar also showed an association between benzodiazepines and the risk of death; and even after adjustment for age, albumin, comorbidities, and other likely confounders, the benzodiazepine–death association remained (Table 2). Although evidence reported recently suggests that symptoms of depression develop as a response to physical deterioration,21.Boulware L.E. Liu Y. Fink N.E. et al.Temporal relation among depression symptoms, cardiovascular disease events, and mortality in end-stage renal disease: contribution of reverse causality.Clin J Am Soc Nephrol. 2006; 1: 496-504Crossref PubMed Scopus (132) Google Scholar we must still be concerned about the dangers of not treating depression and of treating it inappropriately. For example, of great concern is the possibility that underlying the benzodiazepine–death association is a cause–effect relationship. Benzodiazepine-related deaths might be caused by these drugs’ adverse cognitive and psychomotor effects.17.Mann J.J. Drug therapy: the medical management of depression.N Engl J Med. 2005; 353: 1819-1834Crossref PubMed Scopus (323) Google Scholar, 22.Woods J.H. Katz J.L. Winger G. Benzodiazepines: use, abuse, and consequences.Pharmacol Rev. 1992; 44: 151-347PubMed Google Scholar, 23.Shader R.I. Greenblatt D.J. Drug therapy: use of benzodiazepines in anxiety disorders.N Engl J Med. 1993; 328: 1398-1405Crossref PubMed Scopus (319) Google Scholar We are now planning a case-by-case study of benzodiazepine prescription and causes of death.