Original Article

Potentially Inappropriate Medication Use in Older Adults (≥50 years) Living with Human Immunodeficiency Virus (HIV): A Cross-Sectional Study: Polypharmacy in HIV

Jacqueline M. McMillan, MD, MSc1, 2*, Laura Nino Canon, MD1, Shayna Campbell, BSc, Pharm, MSc2, M. John Gill, MB, ChB1, 2

1Department of Medicine, University of Calgary, Calgary, Alberta, Canada;

2Southern Alberta Clinic, Calgary, Alberta, Canada


Background Polypharmacy in older persons with human immunodeficiency virus (HIV) (PWH) is associated with frailty, falls, cognitive impairment, medication nonadherence, and mortality. We characterized polypharmacy and potentially inappropriate medications (PIMs) in older PWH in Calgary, Canada.

Methods We obtained medication reconciliation data for 951 PWH aged ≥50 years as on February 1, 2020. We defined polypharmacy as ≥5 non-antiretroviral therapy (non-ART) medications and PIMs based on the 2019 Beers criteria. We compared older (≥65 years) and younger (50–64 years) age groups, as well as shorter (<10 years) and longer (≥10 years) duration of HIV infection.

Results The mean number of non-ART medications was greater in the older cohort compared to the younger cohort (8.4 versus 6.7, respectively; P < 0.001). Similarly, those living with HIV for >10 years were taking more non-ART medications (mean value, 6.9 non-ART medications) than those with shorter duration of diagnosed HIV infection ≤10 years (mean value, 6.1 non-ART medications) (P = 0.0168). In all 60% were taking ≥1 PIM, with a mean of 1.6 PIMs per patient. Patients living with known HIV infection for >10 years were at greater risk of PIMs (1.6 PIMs) than those with shorter duration since HIV diagnosis (i.e., ≤10 years; 1.4 PIMs) (P = 0.06). PIMs included nonsteroidal anti-inflammatory drugs, opioids, non-benzodiazepine sedatives, gabapentanoids, benzodiazepines, and antipsychotics.

Conclusion Medication reconciliation strategies are required to minimize potentially inappropriate use of medication in older PWH.


Contexte La polypharmacie chez les personnes âgées infectées par le VIH est associée à la fragilité, aux chutes, aux troubles cognitifs, à la non-adhésion aux médicaments et à la mortalité. Nous avons caractérisé la polypharmacie et les médicaments potentiellement contre-indiqués (MPCI) chez les personnes âgées infectées par le VIH à Calgary (Canada).

Méthodologie Nous avons obtenu les données sur le bilan comparatif des médicaments pour 951 personnes infectées par le VIH de 50 ans et plus au 1er février 2020. Nous avons défini la polypharmacie comme étant la prescription d’au moins cinq médicaments non antirétroviraux (non-ARV) et MPCI sur la base des critères de Beers de 2019. Nous avons comparé un groupe d’âge plus avancé (≥ 65 ans) à un groupe d’âge plus jeune (de 50 à 64 ans), de même qu’une durée d’infection par le VIH plus courte (< 10 ans) à une durée d’infection plus longue (≥ 10 ans).

Résultats Le nombre moyen de médicaments non-ARV est plus élevé dans la cohorte plus âgée que dans la cohorte plus jeune (8,4 contre 6,7; p < 0,001). De même, les personnes infectées par le VIH depuis plus de dix ans prennent plus de médicaments non-ARV (6,9) que celles dont le diagnostic de l’infection remonte à dix ans ou moins (6,1) (P = 0,0168). Soixante pour cent des patients prennent au moins un MPCI, la moyenne étant de 1,6 MPCI par patient. Les patients infectés par le VIH depuis plus de dix ans risquent davantage de prendre des MPCI (1,6 MPCI) que ceux dont le diagnostic remonte à dix ans ou moins (1,4 MPCI) (P = 0,06). Les MPCI sont les AINS, les opioïdes, les sédatifs autres que les benzodiazépines, les gabapentinoïdes, les benzodiazépines et les antipsychotiques.

Conclusion Les stratégies relatives au bilan comparatif des médicaments sont nécessaires pour limiter le plus possible l’utilisation de médicaments potentiellement contre-indiqués chez les personnes âgées infectées par le VIH.

Key words: aging, Human Immunodeficiency Virus (HIV), polypharmacy, potentially inappropriate medications, prescribing

Corresponding Author: Jacqueline M. McMillan:

Submitted: 4 February 2021; Accepted: 5 June 2021; Published: 4 February 2022


All articles published in DPG Open Access journals
This article is distributed under the terms of the Creative Commons Attribution-NonCommercial 4.0 International (CC BY-NC 4.0)(


In 2018, an estimated 36% of the 1.1 million persons with human immunodeficiency virus (HIV) (PWH) in the United States were aged ≥55 years, and 17% of all new HIV diagnoses were aged ≥50.1 It has been reported that more than 70% of PWH aged ≥60 years were taking ≥5 non-antiretroviral therapy (non-ART) medications, and 52% were taking a medication that was potentially inappropriate (according to the Beers criteria 2019 used by the American Geriatrics Society [AGS]).2,3 Several studies in developed countries have documented a high prevalence of polypharmacy (defined as ≥5 non-ART medications) ranging between 14% and 55% among older PWH.4

Persons with HIV are at a particularly increased risk of polypharmacy and potential drug–drug interactions because of the requirement of ART medications and medications to treat comorbidities.2 Compared to younger PWH, individuals aged >50 years are more likely to have potential drug– drug interactions requiring cautious drug use and/or dose adjustments.5 In a HIV outpatient study (HOPS), 7% of 3810 patients were concurrently prescribed contraindicated ART and non-ART medications.6 In case of PWH, for each 10-year increase in age, the likelihood of at least one potential drug–drug interaction increased by 1.7-fold (P < 0.0001).5

At the Southern Alberta HIV Clinic (SAC) in Calgary, Canada, polypharmacy among 1329 PWH was associated with the risk of interruptions in ART, independent of the type and number of daily ART medications.7 Reduction over the past two decades in pill burden from the co-formulations of ART has been balanced by the increased use and pill burden of non-ART drugs.8

The recognized harms of polypharmacy in the general population include nonadherence of medication, adverse drug events, increased risk of geriatric syndromes (such as falls, fractures, and dementia), and mortality.9 At SAC in Canada, polypharmacy has been associated with frailty in PWH aged ≥50 years.10 Frailty is an increasing fear in middle-aged and older PWH,11 and recognition that polypharmacy is associated with frailty provides a potentially modifiable treatment target for reducing risk.10 Despite numerous studies investigating the potential harms of polypharmacy in PWH, few prospective or interventional studies evaluate the potential benefits (or unintended harms) of medication reconciliation, reduction in polypharmacy, prioritization of medications, and avoidance of prescribing cascades in PWH.12

Polypharmacy and inappropriate prescribing are not synonymous conditions. PWH with multiple medical comorbidities must be treated appropriately for each condition, unless life-expectancy, frailty, or time-to- benefit becomes disproportionately stronger considerations. Weighing the risk-to-benefit ratio, potential for adverse reactions, and potential for drug–drug and drug–disease interactions requires a delicate assessment and person-centered approach. We evaluated both polypharmacy and potentially inappropriate prescribing in PWH aged ≥50 years from a large, highly characterized, and geographically defined HIV cohort.13


Setting and sample

All PWH residing in southern Alberta receive exclusive HIV care through SAC. Alberta’s publicly funded healthcare system covers the cost of all ART medications, outpatient clinic visits, laboratory tests, and inpatient hospital care under universal health care. Sociodemographic, clinical, medication, and laboratory data of all SAC patients were collected, updated at every clinical visit, and stored in computerized SAC medical database. All SAC patients aged ≥50 years on February 1, 2020, were included in this study.


On clinical visits, pharmacists update non-ART medications by comparing the patients’ medication histories in the AlbertaNetcare Portal (ANP) and the record of non-ART drugs the patient has been receiving in the SAC medical database. The ANP is a web-based clinical portal that records medication dispensing events of all pharmacies located in Alberta.14 Record of all current oral medications (including both ART and non-ART, and prescribed and over-the-counter medications) on February 1, 2020 was extracted from the SAC clinical database. Medications were classified as potentially inappropriate based on the updated Beers Criteria 2019.3 Classes of potentially inappropriate medications (PIMs) included the following: benzodiazepines, non-benzodiazepine sedative-hypnotics [otherwise known as “Z class drugs,” e.g., zopiclone, zolpidem], antipsychotics, opioids, gabapentanoids [i.e., gabapentin, pregabalin], anticholinergics, nonsteroidal anti-inflammatory drugs [NSAIDs], muscle relaxants, antidepressants, antiplatelets, amphetamines [e.g., methylphenidate], triptans, cough suppressants [containing codeine and pseudoephedrine], cannabinoids, buspirone, pramipexole, and anti-Parkinsonian medications [e.g., levodopa/carbidopa and amantadine]. In addition, patients were categorized for medication use according to their age (50–64 and ≥65 years) and the duration of diagnosed HIV infection (0–10 and >10 years).

Statistical Analysis

Statistical analysis was performed using STATA version 14.2 (StataCorp LLC, College Station, TX). Descriptive statistics were calculated, including mean, standard deviation (SD), range, and median of non-ART and PIMs. Similar descriptive statistics were calculated for sample age, gender, and the duration of diagnosed HIV infection.

Chi-square tests and Student’s t-tests were performed to compare group proportions and mean values, respectively, between the older and younger age groups and the groups of shorter and longer duration of diagnosed HIV infection. The significance level (P) was set at 0.05.


Patients attending SAC voluntarily signed a written informed consent form authorizing the use of confidential administrative data for research purposes. The use of non-nominal demographic and clinical data was approved for research purposes by the University of Calgary Conjoint Heath Research Ethics Board (REB19-1575).


Results were reported according to the REporting of studies Conducted using Observational Routinely collected health Data (RECORD) statement.15


As on February 1, 2020, 951/1930 patients aged ≥50 years had received HIV Care at SAC. The mean age of these patients was 59 years (SD: 7; and range: 50–88 years), with a median age of 58 years. Males comprised 82% of the sample (781/951). The mean time since HIV diagnosis was 17.8 years (SD: 8.7 years; and range: 0–38 years), with a median of 17 years. The majority of patients (80%) were aged 50–64 years (N = 761/951), and 20% were aged ≥65 years (N = 190/951) (Table 1).

Table 1. Patient Demographics

Entire cohort N =50 Age 50–64 years N =760 (80%) Age ≥65 years N =190 (20%) Duration HIV infection ≤10 years N =211 Duration HIV infection >10 years N =729
Mean age (years) (standard deviation [SD]) Age range (years) 59.0 (7.0) 50–88 56.3 (4.1) 70.1 (4.9) 57.7 (6.5) 59.4 (7.1)
Females (%) 169 (17.8) 143 (18.8) 26 (13.7) 41 (19.4) 127 (17.4)
Nadir CD4 cell count (cells/µL) Mean (SD) Range 189 (160) 0–1071 192 (162) 178 (152) 232 (196) 176 (145)
Current CD4 cell count (cells/µL) Mean (SD) Range 822 (356) 0–2821 825 (346) 807 (393) 655 (324) 874 (347)
Mean number of comorbidities Median (interquartile range) Range 6.7 (4.8) 6 (3–9) 0–27 4.7 (4.2) 7.2 (4.8) 7.4 (5.3) 6.5 (4.5)
Duration of diagnosed HIV infection Mean years (S.D.) Range (years) 17.8 (8.7) 0–38 17.3 (8.5) 19.4 (8.9) 6.2 (2.8) 21.1 (6.6)
Currently virally suppressed N (%) 881 (92.8) 702 (92.5) 179 (94.2) 184 (87.2) 690 (94.7)

More than two-thirds of the patients (N = 633; 66.6%) were on an ART regimen comprising an integrase inhibitor (INSTI) and nucleoside reverse transcriptase inhibitor (NRTI) backbone. Nearly one-fifth (N = 181; 19%) were on an ART regimen comprising a non-nucleoside reverse transcriptase inhibitor (NNRTI) and NRTI backbone. In all, 52 patients were on a salvage ART regimen (5.5%), 32 patients (3.4%) were on an INSTI and NNRTI backbone, 28 patients (2.9%) were on no active ART, and 25 patients (2.6%) were on a protease inhibitor and NRTI backbone.

Patients were taking a mean number of 6.7 non-ART medications (range 1–27; and median 6). The older cohort (patients aged ≥65 years) were taking significantly more non-ART medications (mean value of 8.4 non-ART medications) compared to patients aged 50–64-years (mean value of 6.3 non-ART medications) (P < 0.001). In addition, those living with HIV for >10 years were receiving a significantly higher mean number of non-ART medications (6.9 non-ART medications) than those with a shorter duration of diagnosed HIV infection of ≤10 years (mean value of 6.1 non-ART medications) (P = 0.0168). There was no difference between males and females in the number of non-ART medications taken (mean value of 6.7 versus mean value of 7 non-ART medications for males and females, respectively; P = 0.2028).

Patients had a mean value of 6.7 comorbidities (median: +/–4.8, range 0–27). The cohort aged ≥65 years had a mean value of 4.7 (median: +/–2.4) compared to a mean number of 7.2 in patients aged 50–64-years (median: +/–4.8; P < 0.001). Patients with a diagnosed HIV infection for >10 years had a mean value of of 6.5 comorbidities (median: +/–4.5) compared to the patients with a diagnosed HIV infection of ≤10 years who had a mean value of 7.4 comorbidities (+/–5.3) (P = 0.0065).

Nearly 60% of the patients (558/951) were taking at least one PIM, with a mean value of 1.6 PIMs per patient (median: 1; range 0–11). Patients living with diagnosed HIV infection for >10 years were at greater risk of PIM (1.6 PIMs) than those with shorter duration of HIV diagnosis (i.e., ≤10 years; 1.4 PIMs; P = 0.06). There was no difference in the number of PIMs between patients aged 50–64-years (1.6 PIMs) compared to patients aged ≥65 years (1.5 PIMs; P = 0.3408). Males and females did not differ in the number of PIMs (1.5 versus 1.8 PIMs for males and females, respectively; P = 0.0991).

The most frequent drug classes involved in PIMs were NSAIDs (24%), followed by antidepressants (22%), opioids (19%), antiplatelet therapy (14%), non-benzodiazepine sedative-hypnotics (“Z class drugs,” e.g., zopiclone, 12%), gabapentanoids (i.e., gabapentin and pregabalin, 10%), benzodiazepines (10%), and antipsychotics (8%) (Table 2).

Table 2. Proportion of Patients Aged ≥50 Years Taking Each Medication Class as on February 1, 2020

Number of Number of patients
patients (%) taking one or more
Total N = different medications
Medication class 951 of the same class
Non-steroidal anti- 225 (23.7) 203 taking 1
inflammatory drugs 21 taking 2
(NSAIDs) 1 taking 3
Antidepressants 205 (21.6) 150 taking 1
46 taking 2
9 taking 3
Opioids 178 (18.7) 138 taking 1
32 taking 2
7 taking 3
1 taking 4
Antiplatelets 130 (13.7) 110 taking 1
20 taking 2
Non- 109 (11.5) 105 taking 1
benzodiazepine 4 taking 2
sedative hypnotics
(i.e., Z-drug)
Gabapentanoids 99 (10.4) 97 taking 1
2 taking 2
Benzodiazepines 93 (9.8) 80 taking 1
13 taking 2
Antipsychotics 72 (7.6) 55 taking 1
14 taking 2
3 taking 3
Anticholinergics 45 (4.7) 40 taking 1
4 taking 2
1 taking 3
Muscle relaxants 43 (4.5) 43 taking 1
Cannabinoids 43 (4.5) 40 taking 1
3 taking 2
Triptans 14 (1.5) 13 taking 1
1 taking 2
Amphetamines 12 (1.3) 12 taking 1
Cough suppressants 8 (0.8) 8 taking 1
(codeine and
Pramipexole 7 (0.7) 7 taking 1
Buspirone 5 (0.5) 5 taking 1
Levodopa/ 3 (0.3) 3 taking 1

Total number of patients, N = 951.

A review of the medical history of 205 patients taking at least one antidepressant found that 80 patients (39%) had a documented history of an affective mood disorder, psychiatric disorder, or suicide risk, and 125 patients (61%) did not have such a documented history. A total of 178 patients were taking at least one opioid medication. On review of their records, 51% (n = 91) had a documented history of neuropathy, polyneuropathy, radiculopathy, myelopathy, migraine headache, or herpes zoster, leaving 49% (n = 87), who did not have one of these conditions or another indication listed for opioid treatment.

There were notable differences in the use of some PIM classes between the younger and older age cohorts (Table 3). Antiplatelet agents were used far more commonly in the older age cohort than in the younger age cohort, 26% versus 10% (P < 0.001); while antipsychotics were more commonly used in the younger age cohort (8.5%) than in the older age cohort (3.7%; p = 0.024). A trend close to achieving statistical significance for increased use of NSAID was observed in the younger age cohort (25%) than in the older age cohort (18%; P = 0.058). Amphetamines were only taken by the younger age cohort, and the difference again approached statistical significance (P = 0.082).

Table 3. Proportion of Patients Taking Each Class of Medication by Age Category, 50–64 years versus ≥65 years of age

Age 50–64- Age ≥65
years years
N =761 N =190
Medication class N (%) N (%) P-value
Non-steroidal anti- 190 (25) 35 (18.4) 0.058
inflammatory drugs
Antidepressants 172 (22.6) 33 (17.4) 0.117
Opioids 141 (18.5) 37 (19.5) 0.765
Antiplatelets 80 (10.5) 50 (26.3) <0.001
Non-benzodiazepine 78 (10.2) 23 (12.1) 0.756
sedative hypnotics (i.e.,
Gabapentanoids 74 (9.7) 25 (13.2) 0.166
Benzodiazepines 77 (10.1) 16 (8.4) 0.481
Antipsychotics 65 (8.5) 7 (3.7) 0.024
Anticholinergics 37 (4.9) 8 (4.2) 0.705
Muscle relaxants 38 (5) 5 (2.6) 0.161
Cannabinoids 37 (4.9) 6 (3.2) 0.312
Triptans 12 (1.6) 2 (1.1) 0.591
Amphetamines 12 (1.6) 0 0.082
Cough suppressants 7 (0.9) 1 (0.5) 0.595
containing codeine and
Pramipexole 5 (0.7) 2 (1.1) 0.568
Buspirone 5 (0.7) 0 0.263
Levodopa/carbidopa 3 (0.4) 0 0.386

Total number of patients, N = 951.

When comparing patients with a diagnosed HIV infection for >10 versus ≤10 years (Table 4), the only significant difference was in the use of antipsychotics. The cohort with a longer duration since diagnosis (>10 years) had a higher proportion (8.6%) of antipsychotic use than the cohort of patients with shorter duration (0–10 years) of diagnosed HIV infection (3.8%; P = 0.019).

Table 4. Proportion of Patients Taking Each Class of Medication by Duration of Diagnosed HIV infection for >10 years versus ≤10 years

Medication class Duration of HIV >10 years N =729 N (%) Duration of HIV ≤10 years n =211 N (%) P-value
Non-steroidal anti-inflammatory drugs (NSAIDs) 175 (24) 49 (23.2) 0.814
Antidepressants 159 (21.8) 43 (20.4) 0.656
Opioids 145 (19.9) 33 (15.6) 0.165
Antiplatelets 104 (14.3) 26 (12.3) 0.471
Non-benzodiazepine sedative hypnotics (i.e., Z-drug) 86 (11.8) 23 (10.9) 0.720
Gabapentanoids 75 (10.3) 23 (10.9) 0.798
Benzodiazepines 77 (10.6) 15 (7.1) 0.137
Antipsychotics 63 (8.6) 8 (3.8) 0.019
Anticholinergics 38 (5.2) 7 (3.3) 0.256
Muscle relaxants 33 (4.5) 10 (4.7) 0.896
Cannabinoids 30 (4.1) 12 (5.7) 0.330
Triptans 12 (1.6) 2 (0.9) 0.461
Amphetamines 8 (1.1) 4 (1.9) 0.363
Cough suppressants containing codeine and amphetamine 6 (0.8) 2 (0.9) 0.862
Pramipexole 6 (0.8) 1 (0.5) 0.603
Buspirone 5 (0.7) 0 0.228
Levodopa/carbidopa 3 (0.4) 0 0.351

Total number of patients, N = 940.

Alcohol, substance use, and smoking status did not differ by age or diagnosed duration of HIV infection (Table 5). Overall, nearly 50% of individuals reported no alcohol intake, and 42% reported <9 drinks per week for females and <14 drinks per week for males. There was no difference by age (P = 0.611) or duration of diagnosed HIV infection (P = 0.187). Slightly over one-third of the patients reported one or more substance use, the most common being marijuana (73%). Again, there was no difference in reported substance use by either the older or younger age cohort (P = 0.075) or by the duration of diagnosed HIV infection (P = 0.748). Lastly, 57.8% of the patients reported never smoking cigarettes, while 25.5% reported current regular smoking. This did not differ by older versus younger age cohort (P = 0.929) or by the duration of diagnosed HIV infection (P = 0.820).

Table 5. Alcohol, Substance Use and Smoking Status, Overall and by Age and Duration of Diagnosed HIV Infection.

Total Age Age Duration of known Duration of known
N (%) ≥65 years 50–64 years HIV >10 years HIV ≤10 years
N =951 N =190 N =761 N =729 N =211
Alcohol use
N (%)
Binge 5 (0.5) 2 (1.1) 3 (0.4) 4 (0.5) 1 (0.5)
Level I 401 (42.2) 74 (38.9) 327 (43) 321 (44) 74 (35.1)
Level II 55 (5.8) 10 (5.3) 45 (5.9) 43 (5.9) 12 (5.7)
Nil 462 (48.6) 97 (51.5) 365 (48) 339 (46.5) 118 (55.9)
Missing 28 (2.9) 7 (3.7) 21 (2.8) 22 (3) 6 (2.8)
Substance use
Total, N (%) 345 (36.3) 57 (30) 288 (37.8) 262 (35.9) 78 (37)
Cocaine 21 (6.1) 1 (1.8) 20 (6.9) 15 (5.7) 6 (7.7)
Crack 16 (4.6) 1 (1.8) 15 (5.2) 13 (5) 3 (3.8)
Ecstasy 2 (0.6) 0 2 (0.7) 2 (0.8) 0
Fentanyl 4 (1.2) 1 (1.8) 3 (1.0) 2 (0.8) 2 (2.6)
Marijuana 252 (73) 41 (71.2) 211 (73.3) 192 (73.3) 56 (71.2)
Methamphetamine 27 (7.8) 10 (17.5) 17 (5.9) 19 (7.3) 7 (9)
Multiple drugs 4 (1.2) 1 (1.8) 3 (1.0) 4 (1.5) 0
Other 19 (5.5) 2 (3.5) 17 (5.9) 15 (5.7) 4 (5.1)
Smoking status
Total, N (%) 951 190 761 729 211
Current Occasional 23 (2.4) 4 (2.1) 19 (2.5) 20 (2.7) 3 (1.4)
Current regular 242 (25.5) 44 (23.2) 198 (26) 181 (24.8) 55 (26.1)
Former >3/12 109 (11.5) 23 (12.1) 86 (11.3) 85 (11.7) 22 (10.4)
Never 550 (57.8) 113 (59.5) 437 (57.4) 422 (57.9) 125 (59.2)
Missing 27 (2.8) 6 (3.2) 21 (2.8) 21 (2.9) 6 (2.8)

Level-I alcohol use: less than 9 drinks per week for females, or less than 14 drinks per week for males.

Level-II alcohol use: more than 9 drinks per week for females and more than 14 drinks per week for males.


In this cross-sectional study of nearly 1000 adult PWH, patients took an average of 6.7 non-ART medications, with the majority taking at least one PIM. Persons living with a diagnosed HIV infection for >10 years were taking more non-ART medications (6.9) than those living with a diagnosed HIV infection for ≤10 years (6.1; P = 0.0168). This could be due to the younger age of onset of HIV infection and higher prevalence of age-related comorbidities seen in PWH compared to persons without HIV;16 future study is required to examine this. In the present study, the cohort aged ≥65 years and living with diagnosed HIV for >10 years had fewer comorbidities than patients aged 50–64 years or with diagnosed HIV infection for ≤10 years. No difference was observed between males and females in either the number of non-ARTs or PIMs, unlike a recent study where females were at 1.75-fold increased risk of taking a PIM.17 The number of non-ART medications and PIMs taken by patients at SAC was similar to the PWH cohort aged >60 years who took a median of 6 non-ART medications.18

The most frequent PIMs we identified were NSAIDs, antidepressants, opioids, antiplatelets, “Z class drugs,” gabapentanoids, benzodiazepines, and antipsychotics. These findings overlap substantially with a recent study of PWH conducted in Spain in which five medications accounted for 85% of PIMs, including lorazepam, ibuprofen, diazepam, metoclopramide, and zolpidem (i.e., two benzodiazepines, one each of “Z class drug” and NSAID).17

Nonsteroidal anti-inflammatory drugs were used by one-quarter of patients. Non-cyclooxygenase-selective NSAIDs are associated with gastrointestinal bleeding or peptic ulcers in adults aged >75 years, and patients taking concomitant corticosteroids, anticoagulants, or antiplatelets.3 For this reason, as well as the risk of increased blood pressure and potential for kidney injury, NSAIDs are advised to be avoided in older adults according to the AGS 2019 Beers Criteria.3 This is particularly important in patients taking concurrent tenofovir disoproxil fumarate (TDF), which is associated with acute kidney injury.19

Although antidepressants were included as potentially inappropriate medications, with nearly 20% of patients taking at least one antidepressant, it is recognized that prevalence of depression could be as high as 37% in PWH (twice that of the general population), with prevalence of up to 70% of PWH who inject drugs.20 Depression has been associated with nonadherence to ART21; hence, an appropriate antidepressant that minimizes adverse reactions and drug– drug interactions must be chosen. There were 205 patients taking at least one antidepressant; of whom, 80 (39%) patients had a documented history of an affective mood disorder, psychiatric disorder, or suicide risk, and 125 (61%) patients did not have such a written history. Among the potential explanations for the 61% without a documented indication for an antidepressant are failure to capture the diagnosis in the medical record, underdiagnosis, and use of an antidepressant for an indication other than a mood disorder (e.g., pain).

Opioids must be used with caution in older adults with a history of, or at a risk of, falls and fractures,3 yet 20% of older adults were taking at least one opioid. The medical records of the 178 patients prescribed opioids demonstrated that 51% of the patients had an indication for opioids, while 49% had no such indication. This again could be due to failure to capture one of these diagnoses in the medical record or use for another indication that was not recorded. Drug levels of many opioids, benzodiazepines, and “Z class drugs” could be increased through drug–drug interactions with boosted ART regimens (e.g., ritonavir and cobicistat).5 Further, in case of PWH taking both opioids and benzodiazepines for long term (≥90 days), there is a nearly two-fold hazard of death (hazard ratio: 1.82; 95% Confidence Interval (CI) 1.27, 2.62).22 At SAC, 10% of older patients were taking benzodiazepines. Benzodiazepines increase the risk of cognitive impairment, delirium, falls, fractures, and motor vehicle crashes in older adults.3 Benzodiazepines may be indicated for individuals with a seizure disorder, rapid eye movement sleep behavior disorder, benzodiazepine or alcohol withdrawal, or severe generalized anxiety disorder.3

“Z class drugs” such as zopiclone, zolpidem, and zaleplon were used by 12% of older patients. The “Z class drugs” have been associated with adverse events similar to those of benzodiazepines (e.g., delirium, falls, and fractures) and hospitalizations and motor vehicle collisions.3 As with benzodiazepines, the AGS 2019 Beers Criteria lists the “Z class drug” as a medication to be avoided in older adults based on moderate-quality evidence.3 Many benzodiazepines and “Z class drugs” have active metabolites, and patients may experience increased toxicity, such as amnesia and confusion, if these are co-administered with protease inhibitors or zidovudine,23 or boosted ART regimens.24

A higher number of patients who were aged 50–64 years (8.5%) compared to patients aged ≥65 years (3.7%; P = 0.024) were taking antipsychotics, and patients living with diagnosed HIV infection for >10 years were also more likely to be taking an antipsychotic (8.6% versus 3.8%; P = 0.019). Two of the primary indications for use of antipsychotics were schizophrenia and bipolar disorder. A conducted US study of PWH found prevalence of bipolar disorder and schizophrenia to be 7% and 6%, respectively, in a cohort of more than 27,000 predominantly male (98%) veterans.25 Our findings of prevalence of antipsychotic use between 3.7% and 8.6% aligns with the prevalence of these mental health diagnoses. However, this does not explain the more frequent use in younger PWH or those living with HIV for >10 years in our cohort, which requires further study.

Many antipsychotics (specifically, quetiapine and olanzapine) may interact with particular ARTs.26 Antipsychotics are associated with an increased risk of stroke, cognitive decline, and mortality in patients with dementia.3 In dementia patients with responsive behaviors, nonpharmacological management is the first line of treatment, and antipsychotics are to be reserved for situations not responsive to nonpharmacological approaches if the individual threatens to inflict harm to self or others.3 In PWH, use of psychotropic medication has been associated with baseline slow gait speed (odds ratio (OR) = 1.61; 95% CI 1.23, 2.10) and increased risk of neurocognitive impairment (hazard ratio: 2.18; 95% CI 1.23, 3.84).27

Future Directions

Polypharmacy in older PWH must be addressed with complete medication reconciliation, ranking of medications according to risks and benefits, and creating a separate plan for each patient.9 A guideline for deprescribing in older PWH recommends regular review of medications, discontinuation of non-indicated medications, prescribing medications only if there is any indication, limit the number of prescribers, and avoid prescribing “cascades.”28 Patients to be prioritized for deprescribing must include those older (i.e., >50 years), frail, having limited life expectancy, and/or are taking multiple medications.28 Deprescribing ART medications has now been an option for PWH as well. PWH having undetectable viral load and a reasonable ART resistance profile could be considered for ART deprescribing, which may include modification to a single-tablet regimen, a two-drug regimen, or a long-acting (injectable) ART.29

It remains imperative that older PWH must not be underprescribed or undertreated for their medical comorbidities. However, polypharmacy is not synonymous with inappropriate prescribing of medication. Approach to the reconciliation of medication of older PWH must be patient-centered and ensure that patients are prescribed indicated medications to minimize adverse reactions and avoid drug–drug and drug–disease interactions. One approach to polypharmacy could include regular pharmacist-led medication reconciliation for older PWH who are taking ≥11 non-ART medications,12 frail,10 or who are nearing the endof-life stage, when the time to benefit from medication may be beyond their life expectancy.


This cross-sectional study ascertained medication history from the SAC medical database. Use of medication was compiled through patient reporting, which is subject to error. For example, patients may use a medication class for a short period and not update the clinic of discontinuation, or may not take the prescribed medication. Additionally, individuals may use over-the-counter medications or out-of-date prescriptions not captured in the electronic database. The inherent weaknesses of self-reporting by patients are improved by pharmacist-led medication reconciliation and review of provincial electronic medical record within Alberta.


In a well-defined cohort of older PWH in Southern Alberta, we found that patients were taking on average six non-ART medications, and 60% of patients were taking a potentially inappropriate medication not based on prescribing guidelines for older adults.3 SAC has implemented routine frailty screening in PWH aged ≥50 years, followed by medication review in frail individuals. However, further prospective and interventional studies are required that could optimize the care of older PWH.


The authors thank the patients and staff at Southern Alberta Clinic, without whom it was not possible to conduct this work. Specifically, the authors thank Quang Vu, data analyst, Jeff Kapler, clinical pharmacist, and Janet Furseth, registered nurse.

Author Contributions

Jacqueline M. McMillan conceived of the idea, performed data analysis, and prepared the manuscript. Laura Nino Canon conceived the idea, determined potentially inappropriate medications, and contributed to the final version of the manuscript. Shayna Campbell conceived the idea and contributed to the final version of the manuscript. Finally, M. John Gill conceived the idea, contributed the data and to the final version of the manuscript.


This work was supported in part by the Brenda Strafford Foundation Chair in Geriatric Medicine, University of Calgary, Alberta.

Conflicts of Interest

Jacqueline M. McMillan, Laura Nino Canon, and Shayna Campbell declare no conflicts of interest. M. John Gill has served as an ad hoc member on the National HIV advisory boards of Viihealth, Gilead, and Merck.


1. Centers for Disease Control and Prevention. Estimated HIV incidence and prevalence in the United States, 2014–2018. HIV Surveillance Supplemental Report. 2020;25(1) ed. 2020. Available from:

2. Greene M, Steinman MA, McNicholl IR, Valcour V. Polypharmacy, drug-drug interactions, and potentially inappropriate medications in older adults with human immunodeficiency virus infection. J Am Geriatr Soc. 2014;62(3):447–53. 10.1111/jgs.12695

3. American Geriatrics Society. Updated AGS Beers criteria for potentially inappropriate medication use in older adults. J Am Geriatr Soc. 2019;67(4):674–94. 10.1111/jgs.15767

4. Freedman SF, Johnston C, Faragon JJ, et al. Older HIV-infected adults: Complex patients (III)-polypharmacy. Eur Geriatr Med. 2019;10:199–211. 10.1007/s41999-019-00195-z 10.1007/s41999-018-0139-y

5. Tseng A, Szadkowski L, Walmsley S, et al. Association of age with polypharmacy and risk of drug interactions with antiretroviral medications in HIV-positive patients. Ann Pharmacother. 2013;47(11):1429–39. 10.1177/1060028013504075

6. Holtzman C, Armon C, Tedaldi E, et al. Polypharmacy and risk of antiretroviral drug interactions among the aging HIV-infected population. J Global Info Mang (JGIM). 2013;28(10):1302–10. 10.1007/s11606-013-2449-6

7. Krentz HB, Gill MJ. The impact of non-antiretroviral poly-pharmacy on the continuity of antiretroviral therapy. AIDS Patient Care STDs. 2016;30(1):11–7. 10.1089/apc.2015.0199

8. Krentz HB, Cosman I, Lee K, et al. Pill burden in HIV infection: 20 years of experience. Antiviral Ther. 2012;17:833–40. 10.3851/IMP2076

9. Edelman EJ, Gordon KS, Glover J, et al. The next therapeutic challenge in HIV: Polypharmacy. Drugs Aging. 2013;30:613–28. 10.1007/s40266-013-0093-9

10. McMillan JM, Gill MJ, Power C, et al. Comorbidities in older persons with controlled HIV infection: Correlations with frailty index subtypes. AIDS Patient Care STDs. 2020 Jul;34(7):284–94. 10.1089/apc.2020.0051.

11. McMillan JM, Krentz HB, Gill MJ, et al. An emerging concern —High rates of frailty among middle-aged and older individuals living with HIV. Can Geriatr J. 2019;22(4):190–8. 10.5770/cgj.22.387

12. McNicholl IR, Gandhi M, Hare CB, et al. A pharmacist-led program to evaluate and reduce polypharmacy and potentially inappropriate prescribing in older HIV-positive patients. Pharmacotherapy. 2017;37(12):1498–506. 10.1002/phar.2043

13. Hanhoff N, Vu Q, Lang R, et al. Impact of three decades of antiretroviral therapy in a longitudinal population cohort study. Antiviral Ther. 2019;24:153–65. 10.3851/IMP3287

14. Alberta Health Services. An overview of Alberta’s electronic health record information system (EHRIS). Updated ed. Edmonton, Canada: Alberta HS; 2015, 64 p.

15. Benchimol E, Smeeth L, Guttman A, et al. The Reporting of studies Conducted using Observational Routinely-collected health Data (RECORD) statement. Plos Med. 2015;12(e1001885). 10.1371/journal.pmed.1001885

16. McMillan JM, Krentz HB, Gill MJ, et al. Managing HIV infection in patients older than 50 years. Can Med Ass J (CMAJ). 2018;190(42):E1253–8. 10.1503/cmaj.171409

17. Lopew-Centeno B, Badenes-Olmedo C, Mataix-Sanjuan A, et al. Potentially inappropriate medications in older adults living with HIV. HIV Med. 2020;21:541–6. 10.1111/hiv.12883

18. Greene M, Steinman MA, McNicholl IR, et al. Polypharmacy, drug-drug interactions, and potentially inappropriate medications in older adults with Human Immunodeficiency Virus infection. J Am Geriatr Soc. 2014;62:447–53. 10.1111/jgs.12695

19. Fernandez-Fernandez B, Montoya-Ferrer A, Sanz AB, et al. Tenofovir nephrotoxicity: 2011 update. AIDS Res Treat. 2011;2011:354908. 10.1155/2011/354908

20. Wagner GJ, Goggin K, Remien RH. A closer look at depression and its relationship to HIV antiretroviral adherence. Ann Behav Med. 2011;42(3):352–60. 10.1007/s12160-011-9295-8

21. Watkins CC, Pieper AA, Treisman GJ. Safety considerations in drug treatment of depression in HIV-positive patients: An updated review. Drug Saf. 2011;34(8):623–39. 10.2165/11592070-000000000-00000

22. Weisberg DF, Gordon KS, Barry DT, et al. Long-term prescription of opioids and/or benzodiazepines and mortality among HIV-infected and uninfected patients. J Acquir Immune Defic Synd. 2015;69:223–33. 10.1097/QAI.0000000000000591

23. Gallego L, Barreiro P, Lopez-Ibor JJ. Psychopharmacological treatments in HIV patients under antiretroviral therapy. AIDS Rev. 2012;14:101–11.

24. Goodlet KJ, Zmarlicka MT, Peckham AM. Drug-drug interactions and clinical considerations with co-administration of antiretrovirals and psychotropic drugs. CNS Spectrums. 2019;24:287–312. 10.1017/S109285291800113X

25. Nurutdinova D, Chrusciel T, Zeringue A, et al. Mental health disorders and the risk of AIDS-defining illness and death in HIV-infected veterans. AIDS. 2012;26:229–34. 10.1097/QAD.0b013e32834e1404

26. Tseng A, Foisy M. Important drug-drug interactions in HIV-infected persons on antiretroviral therapy: An update on new interactions between HIV and non-HIV drugs. Curr Infect Dis Rep. 2012;14:67–82. 10.1007/s11908-011-0229-1

27. Mathur S, Robertson-Toler C, Tassiopoulos K, et al. Detrimental effects of psychotropic medications differ by sex in aging people with HIV. J Acquir Immune Defic Synd. 2019;82:88–95. 10.1097/QAI.0000000000002100

28. Blanco JR, Morillo R, Abril V, et al. Deprescribing of non-antiretroviral therapy in HIV-infected patients. Eur J Clin Pharmacol. 2020;76:305–18. 10.1007/s00228-019-02785-z

29. Guaraldi G, Milic J, Marotullio S, et al. A patient-centred approach to deprescribing antiretroviral therapy in people living with HIV. J Antimicrob Chemother. 2020 Dec;75(12):3425–32. 10.1093/jac/dkaa329