Poor subjective sleep reported by people living with HIV is associated with impaired working memory

Poor sleep can undermine health and may be especially disruptive to those with chronic conditions including HIV infection. Here, clinically well-described people living with HIV [PLWH] (74 men, 35 women) and healthy control (38 men, 35 women) participants were administered the Pittsburgh Sleep Quality Index (PSQI), a validated measure of subjective sleep with a global score ≥5 able to distinguish good from poor sleepers. In addition, participants completed a battery of neuropsychological tests. PLWH (6.8 ± 3.7) had higher global PSQI scores than healthy controls (4.1 ± 2.8): 39.7 % of uninfected controls and 68.8 % of PLWH had a PSQI≥5 indicative of poor sleep. There were no relations between the global PSQI score and any evaluated variables among uninfected individuals or with demographic or HIV-related variables in PLWH. Instead, a higher global PSQI score among PLWH was associated with worse “Quality of Life” scores [Global Assessment of Functioning (GAF, p=0.0007), Medical Outcomes Study survey (21-item short form, SF-21, p<0.0001), and Activities of Daily Living-Instrumental (ADL-I, p=0.0041)] and higher Beck Depression Index (BDI, p<0.0001) depressive symptoms. Further, in PLWH, higher global PSQI scores were associated with poor performance on a working memory task, the digit backward span (p=0.0036). In PLWH, the 5 variables together explained 32.3 % of the global PSQI score variance; only 3 variables – the SF-21, BDI, and digit backward scores – explained 30.6 % of the variance. To the extent that poor subjective sleep contributes to impaired working memory in HIV, we speculate that this impairment may be ameliorated by improved sleep health.

Studies in the general population suggest potential relations between poor sleep and impaired cognitive performance [e.g., 24,25].Among healthy study participants, sleep restriction can induce attention and working memory lapses and depression of mood [26,27].A study using the PSQI in healthy young adults reported that scores ≥5 were associated with worse performance on tests of verbal learning and memory and processing speed [28]; in another study, associations between subjective sleep indices and cognitive performance were not forthcoming [29].In PLWH, relations between poor sleep and cognition are also equivocal [cf., 22, 30-35].In 36 PLWH (75 % men), polysomnography measures (e.g., sleep efficiency, wake after sleep onset), were associated with scores on several neuropsychological tests (e.g., Trails B, Digit Symbol Substitution, Letter Number Sequence) [36].In a similar study of PLWH (n=32), some polysomnography measures of poor sleep (e.g., respiratory disturbance) were associated with low performance on tests of learning and memory (e.g., California Verbal Learning Test), whereas other sleep measures (e.g., sleep maintenance efficiency) were related to poor working memory (e.g., digit span backwards) [37].Comparably, PSQI-determined poor sleep in PLWH has been associated with compromised cognitive performance on the mini mental state exam [38] and tests of learning and memory [39], but studies do not always support relations between PSQI and cognitive performance [31].

Study participants
These data were collected in accordance with protocols approved by the Institutional Review Boards of Stanford University and SRI International.Written informed consent was obtained from all participants in accordance with the Declaration of Helsinki.A total of 182 individuals were administered the PSQI questionnaire between February 2013 and December 2019.The characteristics of the 2 study groups including PLWH (n=109) and healthy controls (n=73) are presented in Table 1.HIV patients were referred from local outpatient or treatment centers, recruited during presentations by project staff at relevant venues, or by distribution of flyers at community events.Comparison participants were recruited from the local community by referrals and flyers.Unrelated data from many of the individuals included in the current study were published in previous reports [54][55][56].

Demographics, clinical screening, and diagnoses
Age, sex, ethnicity, and education were recorded based on self-report.Handedness was measured using the Crovitz Handedness Inventory [57].Socioeconomic status (SES) was derived from the Four-FactorIndex of Social Status, which considers education and occupation level and wherein a lower score reflects higher status [58].Body Mass Index (BMI) was calculated from height and weight measurements made on the day of participation.All participants were screened using the Structured Clinical Interview for DSM-IV and DSM5 (SCID) [59].Upon initial assessment, subjects were excluded if they had a significant history of medical (e.g., epilepsy, stroke, multiple sclerosis, uncontrolled diabetes, or loss of consciousness >30 min), psychiatric (i.e., schizophrenia or bipolar I disorder), or neurological (e.g., Alzheimer's disease) disorders.Participants were excluded for recent (i.e., past 3 months) substance dependence on any drug of abuse other than alcohol.The SCID was used to diagnose Generalized Anxiety Disorder (GAD) and Major Depressive Disorder (MDD) based on established criteria [60,61].During the SCID, the clinical psychologist also assigned a Global Assessment of Functioning (GAF) score -a single rating scale for evaluating the overall functioning of a participant ranging from 1 for the sickest to 100 for the healthiest individuals [62] where scores above 70 indicate positive mental health and scores below 40 are found among hospitalized psychiatric patients [63].The clinical psychologist also determined the Karnofsky score, a tool to assess general well-being with a score of 100 reflecting normal behavior with no signs of disease and 70 indicating an inability to carry on normal activities or to do active work [64].

Questionnaires Pittsburgh Sleep Quality Index [PSQI]:
The PSQI is a 19-item questionnaire that assesses sleep over the past month including quality, latency, duration, efficiency, disturbances, use of hypnotics, and daytime dysfunction (component scores are presented in Table S1).A global PSQI score -which requires completion of all test questions -greater than 5 yields a diagnostic sensitivity of 89.6 % and specificity of 86.5 % in distinguishing good from poor sleepers [20].

Medical Outcomes Study (MOS), 21-item, short form (SF-21):
Health status measures such as the Medical Outcomes Study (MOS) scales have been shown to be useful in HIV infected populations [65,66]; subsets (e.g., SF-21) are sufficient and reliable in determining clinical and functional status in PLWH [67].

Activities of Daily Living (ADL), instrumental and physical:
ADL includes physical activities such as grooming, eating, and using the toilet (9 questions, highest score=18) and instrumental activities such as managing finances and arranging transportation (7 questions, highest score=14) [68].A score of 0 indicates low functioning and high scores indicate independence [69].

The Beck Depression Inventory [BDI]:
The BDI is a 21-item, self-report rating inventory that measures characteristic attitudes and symptoms of depression [70].Each of the 21 items corresponding to a symptom of depression is summed to give a single score with a total score of 0-13 considered minimal, 14-19 mild, 20-28 moderate, and 29-63 indicating severe depression [71].The BDI-II takes approximately 10 min to complete [70] and demonstrates high internal consistency, with alpha coefficients of 0.86 for psychiatric populations and 0.81 for non-psychiatric populations [72].

Alcohol Use Disorders Identification Test [AUDIT]:
The AUDIT was developed by the World Health Organization as a self-report screening Test to identify severity of Alcohol Use Disorders and provides an overall measure of hazardous drinking [73].Hazardous use, dependence symptoms, and harmful use are the three symptom areas covered by the 10-item scale [74].Total scores range from 0 to 40, and higher scores represent more intense drinking [75].

Neuropsychological tests
Executive functioning: Tests of executive functioning included Trails B (time to connect open circles numbered from 1-13 and letters A-L alternating between numbers and letters) [76]; Digit Symbol (the raw number of correct digit-to-symbol substitutions accomplished in 90 sec) [77]; and the F-A-S phonemic fluency test (sum of unique words beginning with the letters "F", "A" and then "S" within 1 min for each letter) [78,79].
Attention and working memory: Tests of attention and working memory included Trails A (time to connect open circles numbered from 1-25); Wechsler Memory Scalerevised (WMS-R) Digit Spans Forward and Backwards raw score totals (verbally repeating -forwards or backwards -a string of numbers spoken aloud by tester) [80]; and WMS-R Visual Spans Block Tapping Forward and Backward raw score totals (participant mimics tapping on a sequence of up to nine identically spatially separated blocks first produced by the examiner) [81].
Visual and verbal learning and memory: Learning and memory tests included the Rey-Osterrieth Complex Figure (ROCF) (immediate and delayed raw scores) and the WMS-R Logical Memory (immediate and delay raw scores).The ROCF evaluates nonverbal memory; participants must reproduce on paper the complex geometrical shape by memory immediately after presentation and following a 30 min delay [82].The WMS-R logical memory subtest is a standardized assessment of narrative episodic memory.The examiner reads aloud a short story, and the participant attempts to recall the story verbatim immediately after hearing the story and again following a 30 min delay [83].

Blood assays
Serum samples were collected and analyzed by Quest Diagnostics for HIV and HCV screening with RNA quantification (viral load) and lymphocyte panel quantification (e.g., CD4 cell count) for seropositive individuals.Other relevant blood markers were also quantified to permit calculation of the Veterans Aging Cohort Study (VACS) index based on age, CD4 cell count, and HIV-1 RNA which predicts mortality and other outcomes in PLWH [84].

Statistics
Statistical analyses were conducted in JMP ® Pro 16.0.0(SAS Institute Inc., Cary, NC, 1989-2021).Data in manuscript are presented are for the global PSQI score; supplementary tables and figures provide similar results using PSQI cutoff scores.All group comparisons used Wilcoxon χ 2 for nonparametric evaluation.Within group correlations used χ 2 for nominal (i.e., categorical) variables and Spearman's ρ for continuous variables.A Bonferroni corrected value of p=0.006 was required for significant correlations (i.e., 10 HIV-related variables, i.e., p=0.05/8).Distinguishing variables were evaluated for their contribution to the global PSQI score using Akaike Information Criterion (AIC), forward stepwise regressions.The variance explaining PSQI global scores after including relevant covariates was derived from standard least squares models.

Group differences
PLWH relative to the control group had a greater number of African Americans, fewer years of education, and lower SES; they also had a high incidence of HCV and were more likely to smoke and drink alcohol (all p<0.0001,Table 1).Further, compared with the control group, PLWH were more like to be diagnosed with GAD or MDD, and had lower GAF and a higher VACS index (all p<0.0001,Table 1).PLWH had significantly higher global PSQI scores (6.8 ± 3.7, χ 2 =25.2, p<0.0001) and higher frequency of poor sleep (PSQI≥5, 68.8 %) than healthy control individuals (4.1 ± 2.8, PSQI≥5, 39.7 %) (Figure 1).Regarding PSQI component scores, PLWH relative to healthy controls had poorer overall subjective sleep quality (component 1, p<0.0001), significantly longer sleep latency (component 2, p=0.0002), and more daytime dysfunction (component 7, p<0.0001) (Table S1).Finally, PLWH relative to healthy controls had worse scores on all administered questionnaires and lower performance on all administered cognitive tests (all p<0.0001, except AUDIT p=0.0247 and trails A p=0.0003, Table S2).

Discussion
The study reported here in 109 PLWH and 73 healthy controls supports poor PSQIdetermined sleep quality in PLWH and extends the literature by demonstrating a salient relation between PSQI and digit backwards performance even after considering relevant variables.Converging results from the PSQI score treated as a continuous variable (i.e., global PSQI score) and as a nominal variable (i.e., PSQI≥5), even after accounting for statistically significant correlates reported in the literature as relevant to PSQI scores in PLWH suggest an effect of poor subjective sleep on working memory in PLWH.Working memory may be among the cognitive functions particularly affected by poor sleep [cf., [85][86][87][88].Poor PSQI-sleep has been associated with low working memory performance in the general population [89,90], in nightshift workers [91], in resident physicians self-reporting memory impairments [92], in bipolar disorder [93], and in schizophrenia [94].
Another significant finding from the current study is the prominent relationship between poor self-reported sleep and worse quality of life in PLWH, comporting with findings frequently reported in the literature [31, 40-43, 52, 95, 96].Here, "quality of life" was assessed using multiple tools including the SF-21, ADL, GAF, and the Karnofsky scores [cf., 43].While the SF-21 and ADL are based on subjective responses to questionnaires (i.e., self-report), the GAF and the Karnofsky are based on a scores assigned by an external, objective observer.Correlations between the PSQI and worse quality of life as measured by the SF-21 and the GAF suggest that both self-reported and externally perceived quality of life are impacted by poor PSQI-defined sleep.
Poor sleep as determined by the PSQI is also often reported in those with MDD [97][98][99][100], but this relation was not forthcoming here.Instead, higher global PSQI scores correlated with higher scores on the BDI, a questionnaire that evaluates depressive symptoms.In several populations, including nurses [101], those with chronic kidney disease [102], alcohol use disorder [98], and multiple sclerosis [103], correlations between PSQI and BDI scores have been reported suggesting that poor sleep can contribute to depressive symptoms.Other reports citing poor PSQI sleep correlates with CD4 count [45][46][47][48][49] or psychiatric diagnoses [e.g., 31,38,52,53] were not replicated.This report also revealed a significant PSQI relation with performance on the Rey-O immediate raw (recall) score, particularly when evaluated using the categorical PSQI≥5 score (note, relative to digits backwards, the Rey-O score contributed less to explaining PSQI variance).This finding comports with papers suggesting high PSQI scores are associated with poor short-term memory [e.g., 90,104,105].
Limitations of the current study include potential bias in the selection of participants contributing to a lack of generalizability of the enrolled sample.Another limitation is a recognized issue with the PSQI as scores do not necessarily correlate with objective sleep measures [22,[106][107][108][109].Nevertheless, the current findings extend the literature by replicating the often-described relationship between poor sleep identified subjectively and compromised quality of life in PLWH and extends the literature by demonstrating a relationship between poor subjective sleep and worse working memory performance after accounting for relevant variables.Indeed, the current results suggest that among all PLWH, sleep is a modifiable disease target that may improve quality of life [110].Further, assessing and treating sleep complaints in HIV might improve working memory functioning.

Supplementary Material
Refer to Web version on PubMed Central for supplementary material.Global Pittsburgh Sleep Quality Index (PSQI) scores in healthy controls and in people living with HIV infection (PLWH).Inset demonstrates percent of total participants in each group with PSQI≥5.NeuroImmune Pharm Ther.Author manuscript; available in PMC 2023 November 09.

Table 1 :
Genaral characteristics of the 2 groups.