Neurology -- Correspondence for Hinkin et al., 59 (12) 1944-1950

Correspondence published:

Reply to Letter to the Editor: Charles Hinkin, SA Castellon, RS Durvasula, DJ Hardy, MN Lam, KI Mason, D Thrasher, MB Goetz and M Stefaniak (14 March 2003)

Medication adherence among HIV+ adults: Effects of cognitive dysfunction and regimen complexity: Adriana Ammassari, Fabrizio Starace, Maria Stella Aloisi, Maria Paola Trotta, Rita Murri, Antonella d'Arminio Monforte, and Andrea Antinori (14 March 2003)

Reply to Letter to the Editor 14 March 2003

Charles Hinkin
UCLA School of Medicine Los Angeles CA,
SA Castellon, RS Durvasula, DJ Hardy, MN Lam, KI Mason, D Thrasher, MB Goetz and M Stefaniak
Send Correspondence to journal:
Re: Reply to Letter to the Editor

chinkin{at}ucla.edu Charles Hinkin, et al.

In the Letter to the Editor describing the Italian Cohort Naīve Antiretrovirals (ICONA) study, Ammassari et al. raised several important points. We agree that neuropsychological dysfunction is likely to be more common among certain subgroups such as older HIV-infected adults and therefore would be a particularly relevant predictor of poor medication adherence among such groups. Also, given the increased prevalence of psychiatric disorder among the HIV-infected population, early diagnosis and treatment of psychiatric illness would no doubt improve medication adherence as well as adherence to other medical directives.
However, we have found no association between depression and antiretroviral adherence when adherence is measured using objective measures. Among 208 HIV+ adults, self-reported depression (using the BDI- II) and 1-month antiretroviral adherence rates, as indexed using computerized monitoring devices, were not correlated (r = .09, p = .28 between log adherence and BDI-II total score). Furthermore, DSM-IV diagnosis of major depressive episode (16% of subjects) also was not associated with poorer adherence. Important methodological differences between the two studies might, at least in part, explain these disparate findings. The Hinkin et al study employed electronic monitoring devices to track adherence, whereas the ICONA group relied upon patient self-report. As Ammassari et al. acknowledge, self-reported adherence often over- estimates actual adherence rates. Indeed, they report that 79% of the ICONA cohort claimed to be perfectly adherent over the preceding week, a figure at considerable variance with most of the existing literature [1].
Our group [2], as well as others [3], has previously demonstrated that data collection method (e.g. self-report vs. objective testing) greatly affects study results. For example, in a separate cohort of 46 HIV -infected adults, we found that self-report of cognitive dysfunction was not associated with actual neurocognitive dysfunction as indexed by objective neuropsychological testing. However, self-reported cognitive dysfunction was associated with self-reported symptoms of depression. Extrapolating from that, we suggest that this phenomenon might explain why the ICONA group found that self-reported adherence was associated with depression, but not with objective neuropsychological testing. The studies that we are aware of that have reported an association between depression and adherence in HIV have typically used self-reported, rather than objective, measures of adherence. Although not a focus of the Hinkin 2002 paper, we conducted additional analyses of self-reported adherence data and found that BDI-2 score was correlated with self-reported number of missed doses (r = .16, p < .05). This finding underscores how assessment method can influence results. We are in agreement with Ammassari et al. in concluding that the causes of poor medication adherence among HIV-infected adults are multifactorial and likely include both psychiatric, as well as neurologic, dysfunction.
References:
1. Chesney MA, Morin M, Sherr L. Adherence to HIV combination therapy. Soc Sci Med 2000; 50:1599-1605.
2. Hinkin CH, van Gorp WG, Satz P, Marcotte T, Durvasula RS, Wood S et al. Actual versus self-reported cognitive dysfunction in HIV-1 infection: memory-metamemory dissociations. J Clin Exp Neuropsychol 1996;18:431-443.
3. Rourke SB, Halman MH, Bassel C. Neurocognitive complaints in HIV- infection and their relationship to depressive symptoms and neuropsychological functioning. J Clin Exp Neuropsychol 1999;21:737-756.

Medication adherence among HIV+ adults: Effects of cognitive dysfunction and regimen complexity 14 March 2003

Adriana Ammassari
Catholic University Rome Italy,
Fabrizio Starace, Maria Stella Aloisi, Maria Paola Trotta, Rita Murri, Antonella d'Arminio Monforte, and Andrea Antinori
Send Correspondence to journal:
Re: Medication adherence among HIV+ adults: Effects of cognitive dysfunction and regimen complexity

aammassari{at}libero.it Adriana Ammassari, et al.

In their study, Hinkin et al. [1] found a greater risk of poor adherence to highly active antiretroviral therapy (HAART) in HIV-infected persons with neuropsychological impairment. Although this finding represents an important step further in HAART adherence research, it might mostly be relevant for some sub-groups of HIV-positive persons among which prevalence of neuropsychological impairment is likely to be higher: those treatment-naive, with AIDS-associated neurologic disorders, or belonging to more advanced age groups. We conducted a study with the purpose to examine the association between depressive symptoms, neurocognitive impairment and HAART adherence. All HAART-treated persons observed between November 1999 and February 2000, enrolled in the Italian Cohort Naīve Antiretrovirals (ICONA) and participating in both the nested studies on adherence (AdICONA) and on neuropsychological function (NeuroICONA) were included in this cross-sectional analysis. Adherence was investigated by a self-administered questionnaire and participants reporting to have missed at least one dose during the last week were considered non-adherent. Neuropsychiatric assessment included the Montgomery-Asberg Depression Rating Scale (MADRS) and the following neuropsychological test battery: motor speed (Timed Gait), fine motor control, selective attention and short-term memory (Digit Symbol subtest), verbal production (Verbal Fluency Test), verbal memory (Auditory Verbal Learning Test, trials 1-5), fine motor control, sustained and selective attention, cognitive flexibility (Color Trails 2). Impairment on each test was defined as a performance >1.5 standard deviations (SDs) worse than the published age -matched norms [2] and global impairment as an impaired performance on at least two out of five tests. Prominent depressive symptomatology was defined as a score >19 on MADRS.
Out of the 135 participants, 36% were women, 31% acquired HIV through drug injection, 42% reported an educational attainment <8 years, 24% were unemployed, 19% had a monthly income of <$350, 9% were currently using drugs, and 10% cited heavy daily alcohol intake. Median age was 35.4 (range, 19-64). Mean value of CD4 was 591x106 cells/l (SD+351) with mean HIV-RNA of 2.5 log10 copies/ml (SD+0.95). Time spent on the last HAART scheme was 0.87 years (mean SD+0.60). Twenty-one percent reported non- adherence, 24% displayed depressive symptoms, and 12% had neurocognitive impairment. Non-adherent participants had higher mean values at the MADRS global score (10.0+8.57 versus 17.5 +9.28 p=0.001). By contrast, there were no significant differences between adherent and non-adherent subjects regarding mean individual neuropsychological test scores and frequency of global neurocognitive impairment. On multivariable logistic regression analysis, non-adherence was associated with worse MADRS score (OR 1.05; 95%CI 1.00-1.10), acquisition of HIV trough drug injection (OR 2.59; 95%CI 1.05-6.39), and complains about impairment of sexual activity (OR 6.62; 95%CI 1.16-37.6).
In this HAART-treated population depressive symptoms, and not neurocognitive impairment, did adversely impact HAART adherence. Explanation for the discrepancy in the results might be found in differences among study-populations: we found younger mean age, higher prevalence of female gender and injection drug users, lower socio-economic status, more frequent neuropsychiatric co-morbidity. Nevertheless, our study might be limited by the use of a self-reported non-adherence which, as already stressed by this article [1], might be a less sensitive measurement among neurocognitive impaired persons. We agree that neuropsychological impairment is likely to be a strong barrier to correct medication intake, but would like to underscore the importance of early identification and treatment of psychiatric co-morbidity as possible intervention strategy to increase HAART adherence.
References:
1) Hinkin CH, Castellon SA, Durvasula RS, et al. Medication adherence among HIV+adults. Effects of cognitive dysfunction and regimen complexity. Neurology 2002;59:1944-1950.
2) Starace F, Baldassarre C, Biancolilli V, et al. Early neuropsychological impairment in HIV-seropositive intravenous drug users: evidence from the Italian Multicentre. Neuropsychological HIV study. Acta Psychiatr Scand 1998;97:132-138.