Abstract
The article, “Associations between depressive symptomatology of neurocognitive impairment in HIV/AIDS” by Tymchuk et al., 1 is a scholarly work that addresses relationships and interactions between depressive symptoms and cognitive impairment in patients with HIV or AIDS. Depressive symptoms have been reported to affect up to 65% of patients with HIV in the era of effective ART, 2,3 and HIV-associated neurocognitive disorder (HAND) is thought to occur at a high rate as well (approximately 50% prevalence across studies and as high as 64% in one study, excluding complaints of cognitive impairment). 4,5 Hence, these disorders may be more common in some samples than cited in the article (45% and 24.5%, respectively). The relationship between these 2 entities is a question of major clinical import today.
The authors state that the overarching aim of this study was to assess the prevalence and severity of depressive symptoms, as assessed by the PHQ-9 scale, in a representative cohort of HIV/AIDS patients in Canada. One of the most important issues not addressed by this article is exactly what is connoted by the presence of “depressive symptomatology.” The “depressive symptomatology” investigated here should not be equated with a diagnosis of “major depressive disorder” (MDD), per DSM-5 criteria. 6 The authors of the paper do not do so. Depressive symptomatology could relate to the presence of other types of depressive disorders, such as bipolar disorder with a depressive episode, cyclothymic disorder, persistent depressive disorder (dysthymia), adjustment disorder with depressed mood, substance-induced depressive disorder, and medication-induced depressive disorder (such as that associated with the antiretroviral medication, efavirenz 7 ). Further, the depressive symptomatology identified here could also relate to none of the foregoing disorders but instead to “no depressive disorder”—simply an elevated depressed mood that can occur at times in anyone in the general population. Of note regarding the differentiation between “symptoms” and “disorder”, it would have been of interest to have the data on the presence of depressive disorders to relate them to another construct on the same level—another disorder (HAND)—rather than relating a type of symptom (depressive) to another disease (HAND) and another construct level (disorder).
The authors report that a subset of 65 patients within the cohort of 265 had a formal psychiatric consultation verified by examining individual patients’ cumulative problem lists. Of those, 24 (a little over one-third) were found to have MDD. However, most of this subset (41 patients; 63%) had other disorders with depressive symptoms or other types of disorders, with anxiety disorders comorbid in 11 patients. Reinforcing an earlier point, the PHQ-9 is appropriately used as a case-finding or screening scale. Elevated scores indicate the need for further assessment but do not indicate a specific diagnosis. Thus, the findings of this study are not sufficiently specific to map to a diagnosis of MDD, which would be the depressive disorder of greatest interest in terms of a relationship to HIV-associated cognitive impairment.
There is also a concern that the symptoms of depression may intrinsically overlap with those of HIV-associated cognitive impairment. For example, cognitive impairment (diminished ability to think or concentrate or indecisiveness) is itself a diagnostic criterion for MDD, 6 and the PHQ-9 8 has an item for “trouble concentrating on things, such as reading the newspaper or watching television.” Hence, both the symptoms of depression and the diagnosis of MDD may be confounded with HIV-associated cognitive impairment and the diagnosis of HAND, as was stipulated by the Frascati Conference criteria cited in the article. 9 However, the association of depressive symptoms with HAND did not persist in a multinomial logistic regression model. It would have been of interest to display all of the data generated by this model. Nevertheless, this result suggests that the connection between these 2 entities may be mediated by, or confounded with, other variables included in the analysis. The complexity of the underlying association is further highlighted by the specific variables that did persist in the multinomial logistic regression analysis: health-related quality of life, sleep problems, and unemployment. Each of these variables could be a cause of depression, or be caused by depression, or both.
It should also be noted that the PHQ-9 was stratified into 3 ranges for statistical analyses according to a widely used grouping: 1) PHQ-9 scores 0 to 4 (interpreted as “minimal depression”), 2) PHQ-9 scores 5 to 9 (“mild depression”), and 3) PHQ-9 scores 10 to 27 (“moderate-to-severe depression”). 8 –11 The article refers to Kroenke et al., 8 who, with regard to syndromal depression, describe the range of scores from 0 to 4 as indicating an “absence of depressive disorder” as opposed to “minimal depression” symptom severity. Scores of 5 to 9 were designated to connote predominantly patients with no depressive disorder but possible sub-clinical symptom levels in a subset rather than “mild depression” symptom severity. Scores from 10 to 27 were designated to connote “a grey zone” from 10-15, with only scores greater than 15 likely to be associated with MDD rather than “moderate-to-severe” depression symptom severity. Along these lines, for a “current major depressive episode”, the PHQ-9 shows a high sensitivity and negative predictive value but a low specificity and positive predictive value. 11 Because MDD has implications for treatment and may share an underlying neuro-inflammatory mechanism with HAND, future studies should consider using diagnostic measures (such as the semi-structured clinical interview for the DSM-5, [SCID]) or, less optimally, single out higher ranges of symptom ratings (such as 15+) to identify patients more likely to have MDD.
In addition, the health-related quality of life measure used in the study was a single-item, 1 to 5 rating of poor/fair/good/very good/excellent. No validity or other psychometric data were cited for this one-item measure, and many studies show that, despite the use of entire scales devoted to measuring health-related quality of life, more work on their validation remains to be done. 12 Quality of life represents a person’s perception of his or her own health status. Depression creates cognitive distortions that lead to negative perceptions of one’s health status. As with sleep disturbances and unemployment, poor quality of life may be a cause or a consequence of depression.
As in other contexts, the role of depressive symptomatology in persons with HIV infection is complex. With regard to HAND status, although it is clear that quantified measurement of neurocognitive impairment and exclusion criteria for other disorders were used to define HAND overall, it was not clear whether functional status criteria were used, which would allow specific levels of HAND severity to be delineated. The conclusion that past and present immunosuppression are not linked to depressive symptomatology should be interpreted with caution, as this study only examined one immune measure (CD4 cell count); yet MDD, for example, has been associated with other immunologic changes, including decreased natural killer cell cytotoxicity, 13 increased activated CD8+ T cell counts, 13 and increased production of pro-inflammatory cytokines. 14 It is also not clear how representative this clinic is of persons with HIV infection living in the region. It may also be noted that the “duration of HIV-1 infection” may not be accurate, as the date of HIV antibody testing does not reflect the date of contraction of HIV infection and may be years separated from it. 15
In summary, the interactions and relationships linking mood and cognition in HIV infection are likely to require determination by a multi-factorial framework and are likely to have bi-directional or higher level elements as well as recursive relationships over time. The study by Tymchuk et al. 1 in this issue of The CJP represents an advance in this area but many questions remain unanswered. Future studies should seek to differentiate depressive symptoms from depressive disorders and should incorporate longitudinal and mechanistic approaches to better understand the causes and effects underpinning these relationships.