Polypharmacy and Potential Drug–Drug Interactions for People With HIV in the UK From the Climate-HIV Database

C Okoli; A Schwenk; M Radford; M Myland; S Taylor; A Darley; J Barnes; A Fox; F Grimson; I Reeves; S Munshi; A Croucher; N Boxall; P Benn; A Paice; J van Wyk; S Khoo


HIV Medicine. 2020;21(8):471-480. 

In This Article


In this large cross-sectional study, we evaluated DDIs in a population of PWHIV in the UK using data from Climate-HIV. The results confirm previous findings, demonstrating that polypharmacy and the risk of DDIs in PWHIV increase with age.[5] This study represents one of the largest, real-world, heterogeneous populations of PWHIV in which DDIs were evaluated. Of the overall population in this study, 2021 patients (44%) were women, 3048 (66%) were non-white, and 1898 (41%) were aged ≥ 50 years. This is in comparison with previous studies evaluating DDIs in PWHIV, which have focused on populations that are predominantly white and male.[5,10]

Most PWHIV are treated with regimens composed of three ARV drugs; therefore, only two additional medications would result in polypharmacy according to the most commonly used definition (≥ 5 drugs).[11] In this study, 47% (2195/4630) of patients were taking at least two non-ARV drugs, and the proportion was higher [61% (1154/1898)] for patients aged ≥ 50 years compared with those aged < 50 years [38% (1041/2732)]. In addition to an increased risk of DDIs, polypharmacy is associated with worst reported outcomes, including elevated risk of adverse drug reactions and non-adherence in the general population and increased risk of geriatric syndromes such as falls, fractures, and dementia in the elderly population.[2,12,13]

In this study, the most prevalent recorded comorbidities were hepatitis B, mental health conditions and hypertension (6% each); this differs from the results of a previous real-world study that indicated much higher prevalence of hepatitis B (24%), mental health conditions (22%), cardiovascular disease (20%), hepatitis C (14%), and diabetes (11%) in PWHIV.[14] In the present study, clinicians recorded patient-reported prescriptions received outside of the participating hospital as well as over-the-counter medications in Climate-HIV; however, as these are patient-reported clinician-entered data, they may not have always been captured reliably, potentially leading to under-reporting of comorbidities. For instance, under-reported use of proton pump inhibitors, which are contraindicated in patients receiving RPV, could have resulted in under-reported DDIs. However, while non-ARV use overall may have been under-estimated, the relative risks of DDIs for different ARV regimens are not expected to be affected, representing a strength of this study.

Despite the possibility of under-reporting, mental health conditions were among the most commonly reported comorbidities, and approximately one in three (32%) of the therapies used for mental health conditions resulted in an amber DDI. Conversely, this observation is consistent with other studies reporting that mental health problems are the most prevalent comorbidities in HIV, with 39–57% of patients having received a mental health diagnosis within the previous 2 years.[15,16] As medications for mental health conditions are most likely to be prescribed by a psychiatrist, they may not appear in medical records for primary care, increasing the risk of inadvertent prescription of interacting drugs. An additional risk that was not recorded in this study is the potential for DDIs when dosages are altered. In this situation, the drug itself remains the same but the change in dosage may lead to a DDI; an alteration may not have been observed until the next clinic appointment with another specialist qualified to amend the dosage or withdraw any interacting drugs.

This analysis found frequent red interactions at the patient level when cross-tabulating the commonly prescribed drug quetiapine with common ARV regimens, though notably, no DDIs are expected between quetiapine and DTG-, RAL-, or RPV-based regimens.[17]

The number of concomitant non-ARV medications tended to increase as patients aged, suggesting an increased presence of comorbid conditions. A previous study also found that comedication use among PWHIV in Switzerland tended to increase with age.[5] Our results indicate that the addition of one non-ARV medication or the presence of any comorbidity or concurrent health condition significantly increased the risk of a red or amber DDI. Overall, regimens based on ATV/r/c or DRV/r/c tended to represent the greatest proportion of patients experiencing red or amber interactions, and the proportion of red and amber interactions also increased with age; regimens based on boosted EVG demonstrated a similar pattern. These regimens were also associated with a greater likelihood of DDI in logistic regression when compared with DTG, as were regimens based on EFV, nevirapine, and RAL (P < 0.05). Dolutegravir and RAL were associated with the highest proportion of green interactions and had no potential red interactions observed.

These results are consistent with previous studies reporting the risks of potential DDIs for PWHIV treated with various ARV regimens. In an analysis from the Swiss HIV Cohort Study, PI- and NNRTI-based therapies were associated with an increased risk of potential DDIs.[5] In a cohort study of PWHIV aged > 65 years treated in France, an increased risk of red DDIs was observed with boosted PIs or EVG (adjusted OR vs. NNRTI-based regimens, 4.12; 95% CI: 3.34–5.10; P < 0.0001) and a decreased risk was observed with DTG- or RAL-based regimens (adjusted OR = 0.02; 95% CI: 0.005–0.050; P < 0.0001).[18] Another recent population-based study evaluated risk of DDIs in PWHIV in a drug-dispensing registry in Madrid, Spain.[19] The greatest risk of red interactions was observed with boosted DRV, while treatment with unboosted INSTI-based regimens was statistically significantly associated with a reduced risk of red interactions (adjusted OR = 0.72; 95% CI: 0.60–0.88; P = 0.001).[19]

The majority of guideline-recommended therapies for treatment of chronic conditions do not consider concurrent conditions. Per the NICE guidelines, the recommended therapies for hepatitis C, hepatitis B, and tuberculosis generated the highest proportions of red DDIs when administered in combination with ARVs. Additionally, recommended therapies for hepatitis C, malignancy, and mental health conditions generated the greatest proportion of amber DDIs, suggesting that these comorbidities require the most interpretation and monitoring by clinicians. However, when reviewing the observed non-ARVs at the patient level, the lipid-lowering medication atorvastatin, the calcium channel blocker amlodipine, and the vasodilator sildenafil were among the most common concomitant medications in this study. These medicines were associated with red, amber or yellow DDIs for most ARV regimens and are all used for the management of cardiovascular disease.

Understanding the classification of DDIs is important for prescribing clinicians, particularly for the red and amber classifications. The red classification designates specific medications that should not be prescribed together. However, the amber classification allows for greater interpretation by the specialist clinician, and the possible dose adjustment and/or monitoring may be associated with greater inter- and intrapatient variability. Therefore, the amber classification may require more interpretation from the multidisciplinary team and better communication with patients. Overall, there were more amber DDI classifications than red or yellow in this study.

This study has several limitations. Data were derived from four selected UK sites and may not be representative of the entire population with HIV in the UK. Concordance between comedications and health conditions was not assessed in this analysis. Additionally, it is unclear whether some potential DDIs identified in this study were managed with dose adaptation, as this would require a longitudinal analysis. Furthermore, we were unable to assess for actual clinical harms resulting from these potential DDIs as data were not collated at a patient level. However, it could be assumed that DDIs classified as red would likely lead to clinical harm. Finally, patients reported prescriptions received outside of the participating hospital along with over-the-counter medications to the clinician at each visit; however, some of these medications may not have been reported as patients may not have been asked by the clinician, patients may have failed to accurately recall them, or prescriptions may have been incorrectly coded when entered into the database. Therefore, medications and comorbidities that were not recorded in the Climate-HIV database may have resulted in misclassification or missing prescription data and under-reporting of DDIs.

This study represents one of the largest, real-world, heterogeneous populations of PWHIV (44% women, 66% non-white) in which DDIs were evaluated. This analysis demonstrated that the use of comedications in PWHIV is high and increases with age, resulting in an elevated risk of polypharmacy and DDIs, which is similar to previous findings from clinical trials and other real-world populations.[5,10] In countries like the UK, most PWHIV on ARVs have undetectable viral loads, see their HIV clinician once or twice annually, and have greater access to medicines over the counter and online. Therefore, efforts to increase communication between healthcare professionals, empower patients and reduce drug burden should be seriously pursued. The results of this study reinforce the importance of ongoing and up-to-date reviews of concomitant medications and comorbidities when seeing a patient in clinic and choosing ARV regimens that can mitigate the associated risks of polypharmacy and DDIs.