After a brief review of epidemiology, the focus is on biochemistry of diabetes. Animal and human studies are reviewed in terms of the impact of alterations in catecholamines and serotonin (5-hydroxytryptamine, 5HT) on glucose utilization. Then, the implications of these experimental results for the choice of antidepressant in comorbid diabetes mellitus and depression as well as in diabetic neuropathy are discussed. Results of clinical investigations are then reviewed in terms of the above hypotheses. An Index Medicus Search for the past 10 years was supplemented by references from previous related reviews of the topic as well as by pending results, where available, not previously published. The range of prevalence of depression in diabetic patients has been 8-27%, depending on study criteria and procedures. An increase of catecholamines appears to increase glucose while both reducing insulin release and reducing sensitivity to insulin that is available. In contrast, increases in serotonergic function by increased precursor, increased release, or blocked metabolism and blocked reuptake in contrast seem to increase sensitivity to insulin and reduce plasma glucose. There have been six studies of fluoxetine, a selective serotonin reuptake inhibitor (SSRI), at a dose of 60 mg/day pursued up to 12 months that have demonstrated that medication's usefulness in diabetic patients, with reductions in weight (to 9.3 kg), in FPG (to 45 mg%), and in HbA1c (to 2.5%). In studies in comorbid diabetes mellitus and depression, nortriptyline, a norepinephrine reuptake inhibitor that produces increased synaptic catechols, has led to worsening of indices of glucose control. However, fluoxetine and sertraline, both selective serotonin reuptake inhibitors, in the same patient group, have produced results consistent with reductions in glucose levels. In diabetic neuropathy, perhaps due to the fact that catecholamines and serotonin may both be implicated in pain pathways, dual-action antidepressants appear more effective at lower doses than do specific serotonergic agents. The tricyclic antidepressants (TCA) (66.7%) have had success in double-blind studies, particularly imipramine, with a 81% response rate. Yet, there are positive reports concerning the SSRIs (paroxetine, citalopram, sertraline), as well as nefazodone, that focus on serotonin selectivity.
Conclusions: In comorbid diabetes mellitus and depression, most evidence supports the use of fluoxetine in control of glucose handling. Other characteristics in terms dosing, drug interactions, cognition, and sleep make sertraline an attractive alternative agent. In diabetic neuropathy without depression, the best choices among non-TCAs may include sertraline, citalopram, and perhaps, venlafaxine, since the TCAs appear to increase cravings and increase FBG levels.