Diabetes mellitus is the most common endocrine disease, with a true frequency of between 1% and 2%.! This illness is distinguished by metabolic abnormalities, ie, elevated plasma glucose levels, which commonly come as a result of either insufficient insulin or resistance to insulin effects. After long-term persistence of elevated blood sugar, frequent complications affect the eyes, kidneys, nerves, and blood levels. This can lead to, among other problems, blindness, kidney failure, and neuropathy. As just stated, diabetes is not a single disease but instead a group of syndromes, which have in common an elevation in glucose. Diagnosis of diabetes mellitus is generally made from a fasting glucose or a laboratory challenge test using oral glucose. In the first case, despite overnight fasting, morning glucose persists on at least two occasions above 140 mg/dL. The second method finds a plasma glucose concentration of greater than 200 mg/dL at 2 hours and on one other occasion after administration of 75 mg of glucose. Two types of diabetes mellitus are commonly described: (I) due to insulin deficiency (IDDM), and (II) because of a reduction in the number of intact insulin receptors (NTDDM). Both conditions may require insulin administration, although type II may alternatively respond to diet alone or diet plus oral hypoglycemics.
DIAGNOSIS OF DEPRESSION IN DIABETES MELLITUS
In terms of the diagnosis of depression in diabetes mellitus, there are three key areas to review: (1) the epidemiology of depression in diabetes, (2) making a proper diagnosis of depression in the diabetic patient, and (3) understanding the impact of depression on the presence and course of diabetes mellitus. In terms of epidemiology, the 19 most recent studies, of which eight were controlled, on prevalence of depression in diabetes were discussed in detail in 1993.2
Nine studies established the diagnosis of depression by structured diagnostic interview. Among these, a prevalence rate of 8.5% to 27.3% was found in the controlled studies, and one of 11.0 to 19.9 in uncontrolled studies. This contrasts to the overall reported community sample rate of 4%.' In these nine studies, the rate found in community samples was 9.6% and 27.3%, and that found in treatment clinics for diabetes mellitus was 8.5% to 22.2%. The severity found of this depression tends to be particularly severe, as discussed in the St. Louis results.3 There, 18 of 28 (64%) of patients had an episode in the previous 12 months. These 28 patients then had a mean rate of relapse of 4.2 episodes in the next 5 years.
The mean age of onset of depression in diabetes mellitus has been found to be 22.1 years in type I diabetes or IDDM and 28.6 years in type II diabetes, or NIDDM.4 This can be compared to the age of onset of depression in the United States, which is typically 27 to 35 years5; thus, it is similar to that found in depression in association with NIDDM. In NIDDM, major depression generally precedes diabetic symptoms; however, in IDDM, the reverse has been shown.3
The impact of depression on the course of diabetes mellitus can be quite severe. For example, a study completed in Germany indicated a direct correlation between severity of depressive symptoms, incidence of complaints of diabetes, and level of hyperglycemia among 128 patients with IDDM.6 The most recent prospective study followed 1715 subjects who were not diabetic at baseline in 1981 from the Baltimore site of the Epidemiologic Catchment Area program.7 Among the 76 subjects who met criteria for major depressive disorder in 1981, 6% or 8% had diabetes on follow up in 1993 to 1994. This contrasted with 80 newly diabetic patients of 1604 (5%) in those not having major depression originally. After controlling for age, sex, race, and socioeconomic status, the risk for NIDDM is revealed to be twice as high in patients with preexisting depression than in those without it. This is consistent with previous results showing significant correlations between Beck Depression Inventory scores and symptoms and signs of diabetic dyscontrol.4 These symptoms and signs with significant correlation, which are possibly a consequence of lack of concentration for selftreatment of diabetes in depression, include both those coming as a consequence of insufficient therapy, leading to hyperglycemia: thirst (0.47) and frequent urination (0.41); and those arriving from too much therapy, leading to hypoglycemia: confused thoughts (0.48), trembling (0.47), sweating (0.37), hunger (0.31), and dizziness (0.25). These reports contrast with those of patients with schizophrenia or bipolar disorder and diabetes mellitus whose psychological illness does not appear to impact on glucose regulation.8 Results presented at the 1996 NCDEU meeting in Boca Raton, Florida, showed that mean HbAIc values (glycosylated hemoglobin) were 9.7 for a combined group of 37 schizophrenic and 16 bipolar diabetics vs 9.6 for an age- and sex-matched group of 53 diabetic control subjects.
Presence of depression in diabetic patients has been found to lead to an increased rate of complications. One study completed at the Los Angeles County University of Southern California Medical Center found the following mean Beck Depression Inventory scores: diabetic patients with complications, 19.1 ± 2.3; diabetic patients without complications, 6.8 ± 0.8; and controls, 5.9 + O.7.9 Severity of depression has been associated in elderly diabetics with worsened glucose control as shown by mean difference of HbAIc from fasting plasma glucose (FPG): low depression: 0.8, medium depression: 0.15, but high depression: -0.25. 10 This group also found the presence of major depression adversely impacts cognitive functioning in the elderly diabetic.11
Making a proper diagnosis of depression in the diabetic patient may be challenging because some symptoms may overlap, eg, lethargy and weight considerations. The Diagnostic and Statistical Manual, 3rd edition-revised and 4th edition (DSM-III-R/-IV)12·13 suggest that these symptoms should only be counted for major depression if they are not "clearly due" to a general medical condition. One resolution14 has included four options: (1) an etiologic approach requiring assessment that a particular symptom or sign is not a result of the physical state, (2) an inclusive approach in which all symptoms are credited without focus on possible physical etiology, (3) an exclusive approach that eliminates any symptoms that may be due to the physical condition, and (4) a substitutive approach of four cognitive symptoms for those physical symptoms being eliminated. Practically, however, this quickly reduces to the exclusive or inclusive approach. The exclusive approach, which is the most pure diagnostically. may indeed be the best for controlled research purposes. However, the inclusive approach, although probably leading to false-positive diagnoses of depression, appears more practical and useful for the clinician because patients, due to stigma, will frequently not report the psychological symptoms of depression. One group has shown, on the basis of Beck Depression Inventory scores, that cognitive symptoms were the most reliable means to separate diabetic depressed from diabetic non-depressed patients.16
BIOCHEMISTRY OF DIABETES MELLITUS
Review of papers from 1966 to 1993 found 15 papers on neurochemistry as related to glucose control.16 In brief, despite the fact that monoamine oxidase inhibitors (MAOIs) are effective as antidepressants because they block the metabolism of norepinephrine and serotonin, their ability to produce hypoglycemia is by a totally different mechanism. The MAOIs act by blocking the long-chain fatty acid oxidation pathway of gluconeogenesis by inhibiting the enzyme carnitine acyltransferase.17 Thus, the hydrazones, with less monoamine oxidase inhibitory affect, produced a greater reduction in glucose than the parent hydrazines, eg, phenelzine.
Catecholamines, by increasing dopaminergic and alpha-adrenergic tone, seem to elevate plasma glucose and reduce insulin secretion by inhibiting glucose-mediated insulin release. For example, the intravenous injection in mice of L-dopa and dopamine led to a 75 to 80 mg/100 mL increase in FBS.18 Norepinephrine injection has been reported to produce 100% blockade of glucose-mediated insulin release.19 Further, the alpha-adrenergic antagonist phentolamine has been found to increase glucose-mediated insulin release.20 The summary of results appears to indicate that an increase in adrenergic tone may lead to insulin resistance.
In contrast, studies of serotonin effects have repeatedly indicated reductions in both plasma insulin and plasma glucose. These results, whether using serotonin precursors (L-tryptophan), serotonin-releasing agents (d,l-fenfluramine and dexfenfluramine), or reuptake blockers (fluoxetine), all appear to be consistent with the principle that increased serotonin tone reduces insulin resistance. For example,21 L-tryptophan (60 mg/kg) reduced plasma glucose from 120 to 70 mg/100 mL without altering plasma insulin. Also, in the presence of 5-hydroxytryptophan, fluoxetine, at 20 mg/kg, and d,lfenfluramine, at 20 mg/kg, both produced significant reductions in plasma glucose in the laboratory. These reductions were, respectively, for fluoxetine, from 6.89 to 2.67 mmol/L, and for d,lfenfluramine, from 6.61 to 4.56 mmol/L.22
ANTIDEPRESSANTS AND EFFECTS ON GLUCOSE IN CLINICAL PRACTICE
As indicated above, MAOIs, due to blockage of long-chain fatty acid oxidation, lead to reduced glucose production. Thus, as early as 1964, their propensity to suddenly reduce plasma glucose to the degree of requiring emergency intravenous glucose was already reported.23 This result has been replicated since. A further problem with the MAOIs is the well-known risk for tyramine-containing foods to lead to sudden increases in blood pressure. It becomes difficult and impractical for a diabetic patient to follow two diets, one for the MAOI and one for the diabetes mellitus. Finally, the propensity of MAOIs to lead to weight gain is well known; this also will lead to complications for the diabetic who needs to control his or her weight.
The tricyclic antidepressants, in general, have predominant effects on norepinephrine over serotonin, eg, in terms of relative selectivity for serotonin over norepinephrine, their ratios are generally less than l.O.24 Sample ratios are: amitriptyline (0.36), Imipramine (0.31), and nortriptyline (0.015). Thus, it would be suspected that these agents would lead to increased insulin resistance and higher plasma glucose. Older case reports and studies presented evidence ifor short-term (6 to 11 days) reduction of fasting blood sugar, followed by longer-term increases by as much as 250%. 16 Most recently, a controlled study of nortriptyline versus placebo was conducted in 79 patients.25 Doses of nortriptyline were controlled double-blind to adjust plasma levels between 50 and 150 ng/mL in these patients with poorly controlled IDDM or NIDDM; approximately 50% also had major depression. It is not surprising that depression scores dropped by twice as much in nortriptyline versus placebo patients, with remission in 57% of nortriptyline patients and only 35.7% of placebo patients. More importantly, though, nortriptyline patients had a net hyperglycemic effect, with worse blood sugar control than placebo patients. In addition, the tricyclic antidepressants are well-known to lead to increases in body weight. This effect can be traced to increased cravings for "sweets" and carbohydrates, which make it harder for the diabetic to follow a special diet.16 For example, patients on maintenance amitriptyline had an 87% rate of significant carbohydrate cravings, contrasted with patients taken off the antidepressant who had an only 29% rate of significant cravings.26 Even worse, this weight gain has been associated with improved response to desipramine in non-diabetics: responders gained a mean of 1.9 lbs whereas nonresponders lost 0.9 lbs after 4 weeks.27 Finally, the effect of tricyclic antidepressants to impair memory, perhaps by their antimuscarinic effect, would tend to worsen the patient's likelihood to follow medical instructions.26
Controlled testing has shown that amitriptyline produces some of the worse results in comparison with other antidepressants in terms of critical reaction time, tracking errors, and so forth, with an overall ranking of 14.8, the worst of a group of 15 antidepressants.28
Serotonin approaches include the serotoninreleasing agents: d,l-fenfluramine and dexfenfluramine, and the selective serotonin reuptake inhibitor (SSRI) class of antidepressants. There are four studies available concerning fenfluramine, one with racemic d,l form (FNF) and three with the dex form (dFNF). As in the laboratory studies above, FNF at a dose of 30 mg twice per day for 4 weeks led to a reduction of 35% in fasting blood sugar without any change in plasma insulin.29 With dFNF, a dosage of 15 mg twice per day for 1 week produced a less significant fall in fasting blood sugar 15%) but also led to reductions of approximately 30% in free fatty acids in obese subjects with NIDDM.30 A 15-day study of dFNF in a double-blind, placebo-controlled study at a dose of 15 mg twice per day in 11 obese female NIDDM patients led to a significant reduction in fasting blood sugar from 120.3 mg/dL to 89.8 mg/dL in dFNF patients, which contrasted with a lack of change in placebo patients.31 Finally, a 12-week double-blind and placebo-controlled study involving 20 obese NIDDM patients reported a significant reduction in HbAIc from 8.5 to 7.1 (P<.05) in the medication group but no change in placebo patients.32
The SSRIs include fluoxetine, sertraline, paroxetine, and fluvoxamine in the United States. All of these medications have a predominant effect for serotonin over norepinephrine reuptake; their respective ratios are: fluoxetine=23, sertraline=64, paroxetine=45, and fluvoxamine=? I.24 Thus, it might be expected that these medications would have effects, like in laboratory studies, to reduce insulin resistance. There have been four studies on the use of fluoxetine in NIDDM patients,3336 but none on its use in depressed diabetic patients. These studies all used doses of 60 mg/day, slightly higher than suggested for initial treatment in major depression (20 mg/day). The first glucose-clamp study indicated that 2-week administration reduced insulin requirements by 20%. 33 The other studies lasted either 6 months34,35 or 1 year.36 The first 6-month study showed a net reduction in HbAIc of fluoxetine over placebo (9.72 vs 10.76; P<.05).34 The 1-year study indicated the presence of fluoxetine-related reductions in weight at each 3-month interval; however, significant reductions in fasting blood sugar and HbAIc were found only after 3 and 6 months.36 The most recent 6-month study, also doubleblind and placebo-controlled, involved 30 obese NIDDM patients over 60 years of age, and found that after both 3 and 6 months, there were significant reductions in both weight and HbA1c with fluoxetine but not with placebo.35 Most recently, an open study of the use of sertraline at a dose of 50 mg/day for 10 weeks was completed in patients with major depression and NIDDM.37 It was found to be effective in significantly reducing Hamilton Depression Ratings scores (P<.001), while at the same time increasing dietary compliance (P<.001). Sertraline also improved HbAIc in over 75% of patients with previously high ratings.
In terms of weight loss associated with SSRI treatment, fluoxetine has reliably led to reductions for up to 6 months of 2.7 to 2.9 lbs, while the others have led to a mean maximum weight loss of only 2.0 lbs.16 However, the reductions produced by fluoxetine have often disappeared after 1 year. There are other considerations, though, in terms of antidepressant choice for a patient with both diabetes mellitus and major depression: pharmacokinetics, sleep, and memory. With regard to pharmacokinetics, factors include half-life of elimination, linearity of kinetics, and drug-drug interactions. Steady-state is established with five half-lives; it is logical that shorter half-lives may lead to shorter time to therapeutic response. Fluoxetine's half-life stands out as much longer, with a norfluoxetine duration of 7 to 10 days; thus, as much as 6 weeks may be necessary to establish efficacy.38 This stands in contrast to both sertraline and paroxetine with a half-life of approximately 1 day.39 Fluvoxamine's half-life of 15 hours is shorter than 1 day and may require multiple daily dosing, which could impact on compliance. Of the four SSRIs, only sertraline has established linear kinetics in usual dosage ranges. Fluvoxamine has a more limited range of linearity; both fluoxetine and paroxetine do not have linear kinetics. This situation can lead to difficulties in trying to predict a response to a particular dose of medication.
In terms of drug interactions, there has been much discussion of the cytochrome P450 microenzyme system.40 Because the diabetic patient may be taking multiple medications, it is important to limit the possibilities for drug interactions. The largest list pertains to 2D6 (IC antiarrhythmics, beta blockers, opiates, tricyclic antidepressants, sparteine, dextromethorphan). Key interactions with other subclasses include 1A2 (eg, theophylline, clozapine), 2C9 (eg, Phenytoin, warfarin, tolbutamide), 2C19 (eg, benzodiazepines, some TCAs), and 3 A4 (eg, carbamazepine, benzodiazepines, anticancer drugs, lidocaine, calcium channel blockers, hormones, erythromycin, astemizole, terfenadine). If a medication blocks metabolism of these agents, a potentially toxic elevation may occur, eg, severe bradycardia from elevations of a beta blocker from 2D6 blockade. Another possible result from blockade is prevention of metabolism of a prodrug into its therapeutic metabolite, eg, 2D6 blockade preventing metabolism of codeine into its therapeutic metabolite.
In terms of 1A2, there is universal agreement that the only major antidepressant to block its effect is fluvoxamine.4043 With regard to 2C9, most reviews40,42,43 focus primarily on fluoxetine with unknown degrees of effect by fluvoxamine (with reported induction of increases of 60% to 98% in plasma levels of warfarin). The other major recent review generalizes the discussion to include sertraline also.41 The most recent reports using phenytoin as the indicator showed significant effects of both fluoxetine and fluvoxamine with lack of effect by both paroxetine and sertraline.44·45 In regard to 2C19, fluvoxamine is reported to have the most effects in contrast to the other SSRIs.43 The biggest controversy has been with regard to 2D6 inhibition, as most reviews40,42,43 have clearly indicated that significantly larger effects of both fluoxetine and paroxetine, one Nemeroff41 has argued for similarities in effect among fluoxetine, paroxetine, and sertraline. However, one review has shown clear laboratory results that indicate significantly greater effects of fluoxetine and paroxetine on 2D6 models of sparteine and dextromethorphan.43 Another, that even at sertraline doses of 150 mg/day, there is less effect on desipramine (72% increase) than with minimal (20 mg/day) doses of either fluoxetine (342%) or paroxetine (364%).42 These results have been replicated with desipramine and dextromethorphan.46,47 Finally, with regard to 3A4, most papers have agreed on the important blocking actions of fluvoxamine and fluoxetine,40,42,43 while another41 additionally indicates a possible role for sertraline. Two reports48,49 have shown lack of impact of sertraline on carbamazepine kinetics, and two others, using a benzodiazepine substrate, have shown significant effects of fluvoxamine and fluoxetine but not sertraline or paroxetine.46,50 Thus, overall, sertraline's profile appears to be least interactive overall with the P450 microenzyme system.
Regarding sleep, all the SSRIs delay the onset of REM (rapid eye movement) or dream sleep and decrease total REM sleep.51 The reuptake blockade of serotonin leads to stimulation of 5HTl which, in turn, produces improvement in depression. However, the simultaneous stimulation of 5HT2 may lead to worsened sleep efficiency and sexual dysfunction.52 Thus, both fluoxetine and paroxetine have been shown to produce increased awakenings during sleep and impaired sleep efficiency.51,53,54 In contrast, perhaps because of its degree of blockade of dopamine reuptake, sertraline has been shown to perhaps reduce awakenings and increase sleep efficiency.53,55
Finally, memory function is important for diabetic patients, who frequently need to stay on a schedule of multiple medications and a special diet. Controlled testing28 of a series of medications showed that only sertraline was significantly better than placebo (rating=4.0). Results on paroxetine were similar to placebo (5.6), whereas those of fluvoxamine (7.6) and fluoxetine (8.6) were worse than placebo. The benefits for sertraline (100 mg/day) over fluoxetine (40 mg/day) in memory improvement in depression were replicated in use of digit symbol and other cognitive tests in a 12-week double-blind study.56 Thus, although all the SSRIs may have a beneficial effect for glucose utilization in diabetics, sertraline may be the medication of choice because of its kinetic safety and benefits for sleep and cognition.
Little is known about the impact of newer antidepressants on parameters of glucose utilization. These include bupropion, venlafaxine, nefazodone, and mirtazapine. In terms of relative specificity to serotonin over norepinephrine, both venlafaxine and nefazodone are more selective to serotonin, while bupropion is relatively devoid of any serotonergic reuptake blockade.24 Mirtazapine does not impact on reuptake blockade; however, at lower doses, it impacts more on release of serotonin than norepinephrine with some reversal of this effect at higher doses.57 According to its focus on adrenergic systems, bupropion might be expected to reduce insulin effects. However, no information is available, but there is contradictory information available concerning bupropion's effects on cravings. There is one report of a generalized increase in carbohydrate cravings.58 However, there are specific case reports of bupropioninduced reductions in cravings for chocolate.59 There are no available reports concerning venlafaxine or mirtazapine. However, regarding nefazodone, there is one recently published case report which is consistent with the above hypothesis that serotonergic selectivity will increase sensitivity to insulin. At a dose of 100 mg bid, a patient experienced two hypoglycemic attacks leading to reduction in the dose of insulin by 15%. w Thus, nefazodone might also be expected to have specific applications for the patient with both major depression and diabetes mellitus; specific trials will hopefully confirm this hypothesis.
Furthermore, in terms of neuropathy, a frequent complication of diabetes, which involves both noradrenergic and serotonergic pathways in pain perception, many antidepressants have been used even in the absence of depression symptoms. There have been 11 prospective studies of tricyclic antidepressants and selective serotonin reuptake inhibitors.61 Amitriptyline and imipramine have been found to be effective in smaller doses, ie, 25 to 100 mg/day, than those used for depression. Both amitriptyline and clomipramine have been more effective than desipramine.62'63 Nortriptyline has been effective at a dose of 20 to 60 mg/day. Analgesic effect requires 1 to 4 weeks. Among the SSRIs, fluoxetine, paroxetine, and sertraline have been studied. In a double-blind trial comparison at a dose of 40 mg/day for 6 weeks, fluoxetine was worse than both amitriptyline (105 mg/day) and desipramine (111 mg/day), with results similar to placebo.62 Paroxetine at the relatively high doses of 30 to 70 mg/day produced analgesia, but over 50% of patients had to drop out because of side effects.64 In another double-blind trial, paroxetine at a dose of 40 mg/day was effective for analgesia but less effective than imipramine.66 Finally, we have results on an initial open study of nondepressed patients with diabetic neuropathy using gradual increases in doses of sertraline to a maximum of 150 mg/day over an 8-week period. Our data has found that as early as week 3, when the dose is at 100 mg/day, there is already a significant reduction in both pain and paresthesias.66
The presence of depression in a patient with diabetes mellitus may significantly worsen the course of both disorders; there may be more depressive relapses and greater numbers of diabetic complications. One needs to pay attention to the antidepressant used for treatment. Noradrenergic approaches, particularly the tricyclic antidepressants, may be effective for depression but by increasing insulin resistance may worsen both fasting blood sugar and HbAIe. In contrast, serotonergic approaches, focusing on the SSRIs, should reduce insulin resistance, therefore leading to both response of depression and reduced HbAIc levels. Fluoxetine at a dose of 60 mg/day has been shown (and the results have been replicated) to be effective in reducing the need for insulin. It has not been studied in major depression with NIDDM. Sertraline, in a preliminary open study, appears to lead to remission of depression and increased dietary compliance and reduced HbAIc levels. Because of the need of diabetics often to take multiple medications and stay on a special diet, the factors of kinetics, sleep, and memory are also important considerations in choice of antidepressant. In these areas, sertraline appears to have the best profile of the SSRIs because it has, eg, linear kinetics and non-impairment of sleep and is the only antidepressant significantly better than placebo on cognition. More work is needed to replicate previous findings and to extend them to other antidepressants.
1. Despodoulos A1 Silvernagi S. Color Atlas of Physiology. New York, NYiThieme Medical Publishers Inc; 1991.
2. Gavard JA, Lustman PJ, Clouse RE. Prevalence of depression in adults with diabetes: an epidemiologic evaluation. Diabetes Care. 1993; 16:1167-1178.
3. Lustman PJ, Griffith LS, Clouse RE. Depression in adults with diabetes. Diabetes Care. 1988; 11:605-612.
4. Lustman PJ, Griffith LS, Gavard JA, Clouse RE. Depression in adults with diabetes. Diabetes Care. 1992; 15:1631-1639.
5. Weissman MM, Leaf PJ, Tischler GL, et al. Affective disorders in five United States communities. Psychol Med. 1988; 18:141-153.
6 . Sachs G, Speiss K, Moser G Glycosylated hemoglobin and diabetes self-monitoring (compliance) in depressed and non-depressed type I diabetic patients. Psychother Psychosom Med Psychol. 1991; 41:306-312.
7. Eaton WW, Armenian H, Gallo J, Pratt L, Ford DE. Depression and risk for onset of type II diabetes. Diabetes Care. 1996; 19:1097-1102.
8. Ellenor G, Marcus D, Ito M. Diabetes control in psychiatric patients. NCDEU Annual Meeting; Boca Raton, FIa; May 1996. Poster 132.
9. Leedom L, Meehan WP, Procci W. Symptoms of depression in patients with type II diabetes mellitus. Psychosomatics. 1991; 32:280-286.
10. Tun PA, Nathan DM, Perlmutter LC. Cognitive and affective disorders in elderly diabetics. Clin Geriatr Med. 1990; 6:731-746.
11. Tun PA, Perlmutter LC, Russo P. Memory self-assessment and performance in aged diabetics and nondi abética. Exp Aging Res. 1987; 13:151-157.
12. Diagnostic and Statistical Manual. 3rd ed, rev. Washington, DC: American Psychiatric Press Ine; 1987.
13. Diagnostic and Statistical Manual. 4th ed. Washington, DC: American Psychiatric Press Ine; 1994.
14. Cohen-Cole SA, Brown FW, McDaniel JS. Diagnostic assessment of depression in the medically ill. In: Stoudemire A, Fogel B, eds. Psychiatric Care of the Medical Patient. New York, NY: Oxford University Press; 1993:53-70.
15. Lustman PJ, Clouse RE, Carney RM, Griffith LS. Characteristics of depression in adults with diabetes. In: Proceedings of the National Institutes of Mental Health Conference on Mental Disorders in General Health Care Settings; Seattle, Wash; 1987; 1:127-129.
16. Goodnick PJ, Henry JH, Buki VMV. Treatment of depression in patients with diabetes mellitus. J Clin Psychiatry. 1995; 56:128-136.
17. Bressler R, Johnson D. New pharmacological approaches to therapy of NIDDM. Diabetes Care. 1992; 15:792-805.
18. Hakanson R, Lundquist I1 Rerup C. On the hyperglycemic effect of dopa and dopamine. Eur J Pharmacol. 1967; 1:114-119.
19. Aleyassine H, Lee SH. Inhibition of insulin release by substrates and inhibitors of monoamine oxidase. Am J Physiol. 1972; 222:565-569.
20. Feldman JM. Inhibition of pancreatic islet monoamine oxidase by adrenergic antagonists and ethanol. Endocr Res. 1975; 2:503-520.
21. Furman BL. The hypoglycemic effect of 5-hydroxytryptophan. Br J Pharmacol. 1974; 50:575-580.
22. Wilson GA, Furman BL. Effects of inhibitors of 5-hydroxytryptemine uptake on plasma glucose and their interactions with 5-hydroxytryptophan in producing hypoglycemia in mice. Eur J Pharmacol. 1982; 78:263-270.
23. Cooper AJ, Keddie KMG. Hypotensive collapse and hypoglycemia after mebanazine, a monoamine oxidase inhibitor. Lancet. 1964; 1:1133-1135,
24. Richelson B. Pharmacology of antidepressants - characteristics of the ideal drug. Mayo Clin Proc. 1994; 69:1069-1081.
25. Lustman PJ, Griffith LS, Clouse RE, et al. Effects of nortriptyline on depression and glycemic control in diabetes: results of a double-blind, placebo-controlled study. Psychosom Med. In press.
26. Paykel ES, Mueller PS, De La Vergne PM. Amitriptyline, weight gain, and carbohydrate cravings: a side effect. Br J Psychiatry. 1973; 123:501-507.
27. Levitt AJ, Joffe RT, Esche I. The effect of desipramine on body weight in depression. J Clin Psychiatry. 1987; 48:2728.
28. Kerr JS, Sherwood N, Hindmarch I. The comparative psychopharmacology of 5HT reuptake inhibitors. Human Psychopharmacology. 1991; 6:313-317.
29. Pestell RG, Crock PA, Ward GM. Fenfluramine increases insulin action in patients with NTDDM. Diabetes Care. 1989; 12:252-258.
30. Sheen AJ, Paolisso G, Salvatore T. Improvement of insulin-induced glucose disposal in obese patients with NIDDM after 1 week treatment with d-fenfluramine. Diabetes Care. 1991; 14:325-332.
31. Greco AV, Mingrone G, Capristo E, De Gaetano A, Ghirlande G, Castagneto M. Effects of dexfenflnramine on free fatty acid turnover and oxidation in obese patients with type 2 diabetes mellitus. Metabolism. 1995; 442(suppl 2):57-61.
32. Willey KA, Molyneaux LM, Yue DK Obese patients with type 2 diabetes poorly controlled by insulin and metformin: effects of adjunctive dexfenfluraniine therapy on glycemic control. Diabet Med. 1994; 11-.701-704.
33. Van Loon BJP, Radder JK, Frolich M. Fluoxetine increases insulin action in obese nondiabetic and in obese noninsulin-dependent diabetic individuals. Int J Obes Relat Metab Disord. 1992; 16:79-85.
34. Gray DS, Fujioka K, Devine W, Bray GA. Fluoxetine treatment of the obese diabetic. Int J Obese Relat Metab Disord. 1992; 16:193-198.
35. Connolly VM, Gallagher A, Kesson CM. A study of fluoxetine in obese elderly patients with type 2 diabetes. Diabet Med. 1995; 12:416-418.
36. O'Kane M, Wiles PG, Wales JK. Fluoxetine in the treatment of obese type 2 diabetic patients. Diabet Med. 1994; 11:105-110.
37. Goodnick PJ, Kumar A, Henry JH, Buki VMV, Goldberg R. Sertraline in co-existing major depression & diabetes mellitue. Psychopharmacol Bull. In press.
38. Rosenbaum JF, Fava M, Quitkin FM, et al. Time to remission of depression with acute fluoxetine treatment. NCDEU Annual Meeting; Marco Island; May 1994. Free Communication 34.
39. Goodnick PJ, Benitez A. New antidepressant agents: recent pharmacological developments leading to increased efficacy. Expert Opinion Investigational Drugs. 1995; 4:935-943.
40. Preskorn SH. Reducing the risk of drug-drug interaction: a goal of rational drug development. J Clin Psychiatry. 1996; 57(suppl 1):3-6.
41. Nemeroff CB, DeVane CL, Pollock BG Newer antidepressants and the cytochrome P450 system. Am J Psychiatry. 1996; 153:311-320.
42. Brosen K. Are pharmacokinetic drug interactions with the SSRIs an issue? Int Clin Psychopharmacol. 1996; (suppl 1):23-27.
43. Riesenman C. Antidepressant drug interactions and the cytochrome P450 system: a critical appraisal. Pharmacotherapy. 1995; 15(6 Pt 2):84S-99S.
44. Schmider J, Greenblatt DJ, von Moltke LL, et al. Inhibition of CYP2C9 by selective serotonin reuptake inhibitors. In vitro studies of phenytoin para-hydroxylation. NCDEU 1996 Annual Meeting; Boca Raton, Fla; May 1996. Abstracts: Poster #122
45. Rapeport WG, Muirhead DC, Williams SA et al. Absence of effect of sertraline on the pharmacokinetics and pharmacodynamics of phenytoin. JCUn Psychiatry. 1996; 57(suppl l):24-28.
46. von Moltke LL, Greenblatt DJ, Court MH, et al. Inhibition of alprazolam and desipramine hydroxylation by paroxetine and fluvoxamine; comparison with other selective serotonin reuptake inhibitor antidepressants. J Clin Psychopharmacol. 1995; 15:125-131.
47. Lam YWF, Ereshefsky L, Riesenman C, et al. In vivo comparison of CYP2D6 inhibition among SSRIs: implications for drug therapy. APA New Research Abstracts 1996 Annual Meeting; New York, NY; May 1996. #NR400, p 176.
48. Preskorn SH. Effects of antidepressants on the cytochrome P450 system. Am J Psychiatry. 1996; 153: 1655-1656.
49. Rapeport WG, Williams SA, Muirhead DC, et al. Absence of a sertraline-mediated effect on the pharmacokinetics and pharmacodynamics of carbamazepine. J Clin Psychiatry. 1996; 57(suppl 1):20-23.
50. Ring BJ, Binkley SN, Roskos L, Wrighton SA. Effect of fluoxetine, norfluoxetine, sertraline, and desmethyl sertraline on human CYP3A catalyzed 1 -hydroxy midazolam formation in vivo. J Pharmacol Exp Ther. 1995; 275:11311135.
51. Sharpley AL, Cowen PJ. Effect of pharmacologic treatments on the sleep of depressed patients. Biol Psychiatry. 1995; 37:85-98.
52. Stahl S. Essential Psychopharmacology. Cambridge & New York: Cambridge University Press; 1996.
53. Winokur A, Reynolds CF HI. The effects of antidepressants on sleep physiology. Primary Psychiatry. 1994; November 22-27.
54. Oswald I, Adam K. Effects of paroxetine on human sleep. Br J Clin Pharmacol. 1986; 22:97-99.
55. Winokur A, Lexow N, Allen K, Reed D, Breitmeyer W. Sertraline administered for 8 weeks to depressed patients did not alter sleep architecture: a preliminary report. Program and abstracts on New Research, American Psychiatric Association Annual Meeting; Philadelphia, Pa; May 1994. NR 212.
56. McEntee W, Oxman T, Ko G, Richter E. The effects of sertraline on cognition in depressed geriatric patients. NCDEU Annual Meeting; Boca Raton, FIa. Poster 65; 1996.
57. Davis R, Wilde MI. Mirtazapine: a review of its pharmacology and therapeutic potential in the management of major depression. CNS Drugs. 1996; 5:389-402.
58. Workman EA, Short DD. Bupropion-induced carbohydrate cravings. Am J Psychiatry. 1992; 149:1407-1408.
59. Mitchell GF, Mebane AH, Billings CK Effect of bupropion on chocolate cravings. Am J Psychiatry. 1989; 146:119120.
60. Warnock JK. Nefiizodone-induced hypoglycemia. Am J Psychiatry. 1997; 154:288-289.
61. Sindrup SH. Antidepressants in the treatment of diabetic neuropathy symptoms. Dan Med Bull. 1994; 41:66-78.
62. Max MB, Lynch SA, Muir J, Shoaf SE, Smoller B, Dubner R. Effects of desipramine, amitriptyline, and fluoxetine in diabetic neuropathy. N Engl J Med. 1992; 326:1250-1256.
63. Sindrup SH, Gram LF, Skjold T, Grodum E, Brosen K, Beck-Nielsen H. Clomipramine vs. desipramine vs. placebo in the treatment of diabetic neuropathy symptoms: a double-blind crossover study. Br J Clin Pharmacol. 1990; 30:683-691.
64. Sindrup SH, Grodum E, Gram LF, Becfc-Nielsen H. Concentration-response relationship in paroxetine treatment of diabetic neuropathy symptoms. A patient-blinded dose-escalation study. Ther Drug Monit. 1991; 13:408414.
65. Sindrup SH, Gram LF, Brosen K, Eshoj O, Mogensen EF. The selective serotonin reuptake inhibitor paroxetine is effective in the treatment of diabetic neuropathy symptoms. Pain. 1990;42:135-144.
66. Jimenez I, Goodnick P, Kumar A. Sertraline treatment of diabetic neuropathy. X World Congress of Psychiatry, Abstracts; Madrid, Spain; August 23-28, 1996, Volume 2, p 313.