DEPRESSION AND DIABETES: UNHEALTHY BEDFELLOWS

Associations between depression and diabetes have been described by physicians for several hundred years. Thomas Willis, an early English physician and anatomist, wrote during the mid-1600s that diabetes is caused by “sadness or long sorrow and other depressions.” In recent years, researchers have posited that there is a bidirectional link between depression and diabetes (Fig. 1).^[1] Depression earlier in life has been shown to be associated with adverse health behaviors, such as lack of physical activity, smoking, and obesity.^[2] Depression also has also been linked to hypothalamic-– pituitary–adrenal -axis (HPA) dysregulation, higher cortisol levels, development of centripetal obesity, lack of insulin sensitivity, higher levels of pro-inflammatory factors, and changes in autonomic nervous system homeostasis.^[3,4] These adverse health behaviors and psychophysiologic changes probably explain the findings from a recent systematic review, which determined that depression in early adult years is associated with an increased risk of development of diabetes.^[5]

Figure 1.

Adverse bidirectional interaction.

Depression has also been linked with poor adherence to self-care regimens among patients with diabetes (i.e. following dietary and exercise regimens, quitting smoking, and taking medications as prescribed).^[6] Both the psychophysiologic effects and the lack of adherence to self-care regimens associated with depression may increase the risk of diabetes complications and mortality. Diabetes complications that are associated with functional impairment may, in turn, precipitate a depressive episode.^[7]

In 2001, we began the NIMH-funded Pathways Epidemiologic Study that screened more than 4,800 patients with predominantly type 2 diabetes enrolled in a large health maintenance organization (HMO). The aims of the Pathways Epidemiologic Study were to: to assess whether depression and diabetes comorbidity was associated with an increased incidence of 5-year all-cause mortality, cardiovascular mortality, incident macrovascular and microvascular complications, and deterioration in activities of daily living; and to assess whether depression and diabetes comorbidity was associated with disparities in quality of health care for clinical control of blood sugar, blood pressure, and lipids. At the time we began the Pathways Epidemiologic Study, several meta-analysis had reported an association of depression with higher HbA_1c levels^[8] and both the number of diabetes complications and each individual complication.^[9] However, the vast majority of these studies were cross-sectional, and, thus, it was not clear whether depression caused poor glycemic control and complications or vice-versa.

The Pathways Epidemiologic Study confirmed previous findings of depression’s effect on mortality by showing that depressed patients with diabetes had more than 50% greater risk of all-cause mortality over a 5-year period.^[10] We had posited an increased risk of cardiovascular mortality among patients with depression and diabetes; however, the risk of cardiovascular mortality was not significant perhaps because of the quality improvements in this HMO and American medicine, in general, in aggressively treating patients with diabetes with cardio-protective medications, such as statins and low- dose aspirin as well as blood pressure control medications.

Diabetes is one of the most common causes of end-stage renal disease. In the Pathways Study, we found that 22% of patients with stage 5 chronic kidney disease (CKD) met criteria for major depression. Over the 5-year period, comorbid depression was associated with a 2-fold increase in mortality rates among patients with diabetes and stage 5 CKD compared to those with diabetes and renal disease alone.^[11] The Pathways Epidemiologic Study data have also shown that over the 5-year period, that patients with comorbid depression and diabetes compared to those with diabetes alone had a 24% increased risk of macrovascular complications, a 36% increased risk of microvascular complications,^{[ 12]} and a 2.7-fold increased risk of development of dementia.^[13] These analyses controlled for socioeconomic, clinical (diabetes duration, HbA_1c, number of diabetes complications, and use of insulin) and health behavior risks (physical inactivity, obesity, and smoking). The fact that these differences were not explained by adverse health behaviors or severity of medical illness emphasizes the importance of examining, in future studies, the neurobiologic mechanisms that could increase the risk of complications, such as depression’s adverse effect on the HPA axis, autonomic nervous system, and proinflammatory factors.

Prior studies among patients with diabetes have shown that poor control of blood sugar, blood pressure, and lipids is explained by both poor adherence to disease control medications and lack of physician intensification of medication regimens (such as adding insulin to a patient’s oral hypoglycemic regimen in patients with poor glycemic control).^[14] The Pathways Epidemiologic Study examined the effect of depression on patient adherence and lack of physician intensification of treatment among diabetics patients with diabetes with poor disease control (HbA_1c≥8.0, systolic blood pressure≥140, LDL≥130) and found that depression was associated with an approximately two-fold higher risk of poor adherence to disease control medications.^[15] However, among those with poor control but adequate adherence, depression was not associated with lack of physician intensification of treatment.^[15]

Although patients with diabetes and depression have significantly worse self-care (adherence to diet, exercise, and cessation of smoking), the high medical utilization of patients with depression and diabetes compared to those with diabetes alone may explain the lack of differences in disease control found in the Pathways Epidemiologic Study.^[16] Although multiple cross-sectional studies had suggested that depression was associated with modestly higher HbA_1c levels in patients with diabetes, we found that those with depression and diabetes compared to patients with diabetes alone only had small, nonsignificant, higher levels of HbA_1c over during a the 5-year period.^[16] We also found no differences over during the 5-year period in control of systolic blood pressure or lipid levels.

We have also examined the possible reciprocal relationships between diabetes and depression by studying whether incident complications of diabetes over during the 5-year period were associated with meeting criteria for major depression at 5 years. We found that incident cardiovascular procedures (such as angioplasty and stent placement) were associated with depression at the 5-year follow-up.^[7] In this paper, we also showed the remarkable chronicity of depression in patients with diabetes. More than 80%of those with major depression at the 5-year follow-up either met criteria for major or minor depression 5 years earlier.^[7]

In addition to the large epidemiologic study described above, we have now completed two randomized trials testing a collaborative depression care intervention versus usual primary care to improve the quality of depression care and depressive outcomes in patients with comorbid depression and diabetes.^[17,18] Both of these trials (Pathways and IMPACT) tested an intervention, which included behavioral activation, a depression care manager who provided a choice of starting treatment with either problem-solving therapy (PST) or antidepressant medication, a stepped care algorithm, and a relapse prevention plan. The stepped care algorithm was developed to provide an increased intensity of care for patients with persistent depressive symptoms despite initial treatment. Thus, if a patient did not respond to an initial choice of antidepressant medication, another medication could be tried or PST could be added and if did he/she did not respond to an initial treatment with PST, medication could be added. These multimodal interventions were shown in both studies to improve the quality of depression care and depressive outcomes over during a 2-year period.^[17,18] Patients in the intervention arm were more satisfied with depression care and 30% more likely to report that they were “very much” to “completely” improved. HbA_1c levels were relatively low (i.e. 7.8%–8.0%) at baseline and did not show significant intervention versus usual care differences over time.

Two cost-effectiveness analysies from these studies showed that the $500 to $700 increased mental health costs associated with the collaborative care interventions were off-set by greater savings in total medical costs over during the 2-year period.^[19,20] Both intervention studies also showed that collaborative care was also associated with significantly more depression-free days over during the 2-year period compared to usual care.^[19,20] A third paper article that we recently published from one of these two studies has shown the same trends in medical cost savings associated with the intervention continued during years 3 to 5.^[21] Thus, enhanced treatment of depression in patients with diabetes appears to place patients on a different trajectory of medical costs for up to 5 years.

The lack of significant effect on HbA_1c despite the beneficial intervention effect on depression outcomes has led us to hypothesize that treatment of depression is a necessary first-step, but not sufficient in itself to affect medical disease control. The relatively low HbA_1c levels in the previous studies also limited the potential beneficial effect of improving depression because depression is only modestly related to higher HbA_1c levels.^[8] These data have led us to the ideas developed for our currently funded NIMH trial in which we have screened patients with diabetes and/or coronary heart disease (CHD) for poor disease control (HbA_1c≥8.5%, systolic blood pressure >140mm Hg, LDL>130) and if ≥1 of these disease control parameters were high, the patients also received the PHQ-9.^[22] We then offer randomization to patients with PHQ-9 scores of ≥10 and evidence of poor control on at least one of the above three disease control measures to a nurse intervention versus usual primary care. The intervention we are testing has focused on both improving depression outcomes (with a combination of behavioral activation and enhanced antidepressant management) and medical disease control. The nurses are supervised weekly by a psychiatrist and primary care physician and bring recommendations about changes in either antidepressant or medical disease control medication to the patient’s primary care physician. Because of the potential adverse reciprocal effects of depression and poor medical illness control, we hypothesized that our intervention verses control effect size on depression outcomes would be greater in this trial and that we would also show significant improvement in medical disease control. Our results from this trial will be reported in 2010.

CONCLUSION

Our epidemiologic research on depression and diabetes has demonstrated the reciprocal links between brain and body. Our intervention research has demonstrated the effectiveness of enhanced screening and provision of collaborative care to patients with comorbid depression and diabetes on improvement in depressive symptoms, functioning, and savings of medical costs. Finally, our new research treatment models will also hopefully show the importance of utilizing models that address both depressive symptoms and poor medical disease control on improving both psychiatric and medical outcomes.

Biography

Wayne Katon, M.D., is Professor of Psychiatry, Director of the Division of Health Services and Epidemiology, and Vice Chair of the Department of Psychiatry and Behavioral Sciences at the University of Washington Medical School. He is Director of a NIMH-funded National Research Service Award Primary Care-Psychiatry Fellowship who has successfully trained psychiatrists and primary care physicians for academic leadership positions. Dr. Katon is internationally renowned for his research on the prevalence of anxiety and depressive disorders in primary care, the relationship of psychiatric disorders to medically unexplained symptoms, such as headache and fatigue, and the impact of depression and anxiety on patients with chronic medical illness. In recent years, his research has focused on developing innovative models of integrating mental health professionals and other allied health personnel into primary care to improve the care of patients with major depression and panic disorder. Dr. Katon has been awarded the American Academy of Family Practice Award for Excellence in Teaching in Primary Care numerous times. He also has also been awarded the Academy of Psychosomatic Medicine Research Award (1993) and the American Psychiatric Association Senior Scholar Health Services Research Award (1999) and the Depression and Bipolar Support Alliance Gerald L. Klerman Senior Investigator Award (2003). He is Editor-in-Chief of General Hospital Psychiatry and is honored by being one of the Web of Science Highly Cited Authors. Dr. Katon has written more than 400 peer-reviewed journal articles and chapters, as well as Panic Disorder in the Medical Setting, a book for primary care physicians. In addition, Dr. Katon and his research team have written a self-help book for depressed patients titled Depression: Self-Care Companion for Better Living.