Patients with diabetes are two times more likely to present with symptoms of depression; how does this affect diabetes outcomes, and is the need for depression screening in patients with type 1 different from those with type 2?

Previous studies have found depression to influence diabetes-related outcomes. These studies have shown mixed results on linking depression and glycemic control. Furthermore, these studies are typically performed on patients with both type 1 and type 2 diabetes, or in patients with type 2 diabetes. There isn’t much evidence for the role of depression or depression screening patients with type 1 diabetes.

This study examined changes of depression in patients with type 1 diabetes over a 4-year span, while also assessing the association between changes in depression and glycemic outcomes. The authors hypothesized that improved depressive symptoms would result in improved glycemic outcomes. Using the Type 1 Diabetes Exchange Clinic Network and registry, 2,744 patients with type 1 diabetes were chosen. Each completed the PHQ-8 at baseline and follow-up, which was 5 years after enrollment. The PHQ-8 was a self-report assessment used by the researchers for depression symptoms and their severity. Major depression was defined by a PHQ-8 score of > 10. To better the generalizability of the general population, a scoring algorithm used in the Behavioral Risk Factor Surveillance Survey (BRFSS) was also used. It was noted that depression was not being diagnosed, but stated as probable major depression, and should be further looked at with a practitioner for follow-up.

Patients were then split into one of four groupings based on their “elevated depression symptoms (EDS).” “Persistent EDS” was defined as having EDS at baseline and follow-up, “Resolved EDS” were those with EDS at baseline, but not at follow-up, “New-Onset EDS” were patients without EDS at baseline, but present at follow-up, and “Non Depressed” were patients without EDS at baseline and follow-up. Demographics were retrieved via questionnaire and medical record data was used for information on age, duration of diabetes, insulin modality, use of CGM, and BMI. HbA1c values were measured 13 days after PHQ testing, then again at follow-up. Any severe events were to be recorded, such as severe hypoglycemia (SH) or diabetic ketoacidosis (DKA) within 3 months from baseline and again at follow-up.

To assess different associations, multiple models were used, such as multivariable linear regression models, linear regression models, separate multivariable logistic regression models, and multinomial logistic regression models. Normally distributed variables were expressed as mean + SD and non normally distributed variables as the median. Covariates were adjusted for, if an association was present, it was included in the model for that outcome. True associations were considered with a  P-value of < 0.01.

Comparing EDS from baseline to follow-up using the PHQ-8 definition, of the 2,744 patients, 131 (5%) had Persistent EDS, 122 (4%) with Resolved EDS, 168 (6%) with New-Onset EDS, and 2,323 (85%) defined as Not Depressed. Results were similar when defining EDS using the BRFSS algorithm, with a slight elevation in the Not Depressed group. An association was found between depression symptom status from baseline to follow-up and patient characteristics. Specifically, the Not Depressed group between baseline and follow-up, were found to be older, more likely to be male, more likely to have private insurance at baseline, and more likely to have started a CGM between baseline and follow-up. Comparing the Persistent EDS and New-Onset EDS group, they were less likely to be using a CGM at baseline or follow-up, and the New-Onset EDS group were younger and more likely to be female.

In each group, the HbA1c values increased from baseline to follow-up, with a smaller increase in the Resolved EDS and Not Depressed groups. The association was seen between change in HbA1c and in depression symptom status from the PHQ-8, BRFSS algorithm, and continuous change in PHQ-8 score, with an adjusted P < 0.001. The researchers determined an increase in the PHQ-8 related to an increase in HbA1c. In regards to SH and DKA occurrence, via the PHQ-8 definition, patients with Persistent EDS and New-Onset EDS were more likely to report SH, versus those with Resolved EDS, were less likely to report SH (adjusted P = 0.11). The BRFSS algorithm had similar results (P = 0.21). But no association was reported between SH at follow-up and continuous change in the PHQ-8 score (P = 0.70). Similar results were also seen for DKA occurrence for depression symptom status for both PHQ-8 and BRFSS algorithm. Persistent EDS and New-Onset EDS groups were more likely to experience DKA at follow-up versus Resolved EDS and Not Depressed groups (adjusted P < 0.001). Furthermore, those with an increase from baseline to follow-up in the continuous PHQ-8 score were more likely to report DKA at follow-up (adjusted P = 0.03). There was no found association between change in BMI z-score and depression symptom status.

The authors believe their results show a connection between depression symptom status and glycemic outcomes over time which warrants further studies. This connection was most prominent in patients with persistent or significant depression symptoms and worsening glycemic control. Furthermore, a similar relationship was noted between depression symptom status and the occurrence of SH and DKA. This all suggests that patients with type 1 diabetes also presenting with depressive symptoms are at an increased risk for both short and long-term complications due to its effect on glycemic levels. While causality can’t be started, the relationship is apparent. In this study, to determine depression status, a self-report was used. While this self-report was validated, there is a chance for misclassification. Part of having a large sample size was in hopes to diminish any effect of misclassification. Also, SH and DKA incidence were self-reported, with no confirmation, also potentially affecting the results. Similarly, for future studies, it might be important to know if depression is treated in the study and if this affects glycemic control, which the authors didn’t have access to for this study.

Practice Pearls:

  • Patients with diabetes are already more likely to have depressive symptoms.
  • Routine depression symptom screening should be implemented in patients with diabetes for better long-term outcomes.
  • Diagnosis and effective treatment theoretically should lead to better long-term diabetes outcomes but future studies are warranted.

Reference for “The Importance of Routine Depression Screening in Patients with Type 1 Diabetes”:

Trief, Paula, et al. Longitudinal Changes in Depression Symptoms and Glycemia in Adults With Type 1 Diabetes. Diabetes Care. 2019 May 21.

Emma Kammerer, L|E|C|O|M Bradenton School of Pharmacy, PharmD Candidate