Beers versus STOPP criteria in polyharmacy community-dwelling older patients

Criterios de Beers versus STOPP en pacientes mayores, polimedicados y residentes en la comunidad

This work is licensed under Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International.

Medication prescription in older patients will often be a complex process, due to the fact that they will usua- lly present multiple comorbidities. Moreover, changes in homeostasis, pharmacokinetics and pharmacodynamics derived of aging will render older patients more prone to suffer adverse reactions to medications, therefore increasing morbidity and the need for medical care¹.

Adverse reactions to medications represent around 7.2-16.8% of admissions of elderly patients in Internal Medicine Units^2,3,4, and in many cases, a potentially inappropriate prescription (PIP) of medications is one of the factors leading to hospitalization^5,6,7,8. This term refers to those situations where the risk to suffer adverse events is superior to clinical benefit, particularly when there are safer and/or more effective therapeutic alternative options. Moreover, the use of drugs with a frequency, duration or dose superior to their indication is also considered inappropriate, as well as the use of medications with a high risk of interactions, and therapeutic duplications. On the other hand, it is also inappropriate to omit the prescription of medications with a clearly demonstrated indication for specific situations^9,10,11.

Implicit and explicit methods have been developed for the assessment of PIP in older patients¹². Implicit methods are based on the clinical evaluation of patients and their medication, and the most widely used is the “Medication Appropriateness Index” (MAI) Method ¹³. Explicit methods measure PIP by means predefined criteria based on evidence and expert consensus. Among these, Beers criteria^14,15,16,17 and “Screening Tool of Older Person´s Prescriptions/Screening Tool to Alert doctors to Right Treatment” (STOPP/START)^18,19 criteria are the most widely used. Beers criteria were developed in U.S. in 1997¹⁴ and updated in 2003¹⁵, 2012¹⁶ and 2015¹⁷. In their 2012 update¹⁶, besides the review of molecules, there was an addition of pharmacological groups, and information was included about the reasons upon which PIP criteria are based. The STOPP/START criteria¹⁸ were published in 2008 in Ireland, due to the limitations of extrapolating Beers criteria to the European setting, and the difficult application of the implicit methods; these have been updated in 2015¹⁹. The STOPP/START criteria have demonstrated high reliability among evaluators, and have been adapted to our setting^20,21. These criteria are focused on potential problems associated with drug treatment, in the setting of the comorbidities inherent to older patients, and they detect not only overprescription and PIP, but also treatment omissions^22,23.

With the aim to assess PIP in > 65-year-old polypharmacy patients and understand its prevalence, we conducted an analysis of prescriptions and electronic clinical records, applying Beers criteria¹⁶ and STOPP criteria¹⁸, in the relevant Primary Care Records from the Healthcare Area II of the Principality of Asturias Health System (SESPA).

A descriptive, transversal study was conducted on those prescriptions written at the health centres from the SESPA Healthcare Area II, for polypharmacy≥ 65-year-old patients. In 2011, this Healthcare Area had a population of 28,808 inhabitants, and 8,117 (28%) of them were 65-year-old or older, according to the annual count of Individual Health Insurance Cards.

The study was conducted based on the information from: 1) invoicing data from the SESPA Pharmacy Department, and 2) electronic clinical records (OMI-AP). Based on the prescription invoice data, lists were drawn for drugs prescribed to ≥ 65-year-old patients, who were taking chronically 10 or more drugs/day during 6 or more conse- cutive months (from October, 2010 to March, 2011). Those drugs with topical administration were excluded from the drug count. For each pharmacological treatment, the lists included the name of the molecule, its brand name, the number of packages sold, and the date of invoicing. This information was analyzed and compared with the information from the electronic clinical history of those patients included in the study. For drug analysis, the national codes and the Anatomical-Therapeutical-Chemical (ATC) Classification codes were manually added to the data matrix. Those medications that included combinations of 2 or more active substances were accounted as the number of active substances included.

In order to assess PIP, the following criteria were used:

a) Beers criteria, independent of diagnosis (ID) and considering diagnosis (CD) (2012 version)¹⁶ and b) STOPP criteria (2008 version)¹⁸, adapted and translated into Spanish by Delgado Silveira et al.20. The assessments applied for some of these criteria appear in table 1. In each of the analyzed cases, the presence of any of Beers criteria and STOPP criteria was studied. There was a calculation of the frequencies and percentages of each of the criteria analyzed. Comparisons were conducted through Chi-Square test, using the IBM SPSS Statistics 22 Program.

The treatments for 223 patients in total were analyzed; 128 (57%) were female, with a median 78-year-old [range] age [65-96]. The median [range] of medications/day taken by the patients in the study was 12 [10-19] medications.

Beers Criteria

In total, 141 (63.2%) patients had at least one PIP according to total Beers criteria (Table 2). The median [range] of Beers criteria per patient observed was 1 [0-6]. Identified Beers criteria ID and CD are collected in Tables 3 and 4. The highest prevalence of Beers criteria was found in those drugs acting at the Central Nervous System level, followed by those used for pain management, and cardiovascular system drugs. Thus, in the Central Nervous System section, 110 (ID) and 48 (CD) patients had at least one Beers criterion; 35 patients presented some ID criterion associated with pain management, and in the cardiovascular system, 30 (ID) and 10 (CD) patients had at least one Beers criterion.

The 3 most frequent Beers Criteria ID (Table 3) were:

1) use of benzodiazepines (ID 5.5), 2) use of non-COX selective non-steroidal anti-inflammatory drugs (ID 8.2), and 3) the use of alpha-1 adrenergic blockers in the cardiovascular system (ID 4.1). Regarding Beers Criteria

Table 1
Assessment applied for some of Beers and STOPP criteria

Table 2
Distribution of the criteria observed in the analyzed sample

Table 3.
Identified Beers Criteria Independent of Diagnosis (ID)

CD (Table 4), the three most frequently observed criteria were: 1) the use of inhaled anticholinergic drugs or strongly anticholinergic drugs, except for urinary antimuscarinic drugs with lower urinary tract symptoms or benign prostatic hyperplasia (CD 4.3), 2), the use of benzodiazepines, antipsychotic drugs, zolpidem or H2 antihistamines, in patients with dementia or cognitive deterioration (CD 2.3), and 3) selective serotonin reuptake inhibitors, benzodiazepines, zolpidem or antipsychotic drugs in patients with a history of falls and fractures (CD 2.4).

Table 4.
Identified Beers Criteria Considering Diagnosis (CD)

STOPP Criteria

Out of all the patients analyzed, 165 (73.9%) had at least one PIP according to STOPP criteria (Table 2). The median [range] of STOPP criteria observed was 1 [0-6]. The drugs involved affected more frequently the cardiovascular and musculoskeletal system, and the central nervous system (Table 5). Therapeutic duplication, longterm use of benzodiazepines with long half-life, and the use of acetylsalicylic acid at doses> 150mg/day were the 3 STOPP criteria most frequently observed. The most frequent duplications were associated with the use of antiaggregant drugs, non-steroidal anti-inflammatory drugs, benzodiazepines, and antidepressants.

The prevalence of PIP observed in our study, by applying both STOPP and Beers criteria, was high compa- red with some previous publications^{21,23,24,25,26,27,28}. However, the number of medications received seems to be directly related with PIP frequency²⁹. This is probably the reason why PIP rates are usually superior in hospitalized patients^{21,30,31,32,33}vs. patients seen by Primary Care^21,24,34. Our study is focused on the analysis of PIP in patients with a high polypharmacy level, with 12 [10-19] medications/ day, which might explain the high prevalence observed. Moreover, we have detected PIP most frequently when applying STOPP criteria than Beers criteria, which coincides with a great part of previous observations ^23,35,36. Thus, an European study published in 2011³⁷, which analyzed PIP in hospitals located in Switzerland, Spain, Belgium, Italy, Czech Republic and Ireland, found in all countries a higher number of PIP with STOPP criteria than with Beers criteria. Subsequently, Blanco-Reina et al.³⁴ have analyzed PIP in outpatients, using Beers criteria (2003 version), STOPP criteria, and Beers criteria (2012 version); the PIP rates observed were 24.3%, 35.4% and 44%, respectively. However, our study, whe-

Table 5
STOPP criteria identified with a≤ 6 frequency

re the 2012 version of Beers criteria was used, detectedhigher PIP rates than with STOPP criteria. In our opinion,these differences could be based on differences in thesample of patients analyzed. Specifically, patients in ourstudy were taking at least 10 medications/day, while inthe study by Blanco-Reina et al.³⁴, the mean number ofmedications was 4.5 ± 2.9. This also coincides with thefact that the STOPP criteria most frequently observed inour study was medication duplication, while in the studyby Blanco-Reina et al.³⁴ this was the 5th most frequentSTOPP criteria.

The three pharmacological groups mostly associated with PIP in our study were benzodiazepines, antiaggregant drugs, and non-steroidal anti-inflammatory drugs.

By applying the STOPP criteria we found, similarly to some previous studies^24,27, a high PIP with drugs involving the cardiovascular system, followed by those involving the central nervous system; while when applying Beers criteria, we observed that the first place was for PIP of drugs involving the central nervous system, maybe because these criteria include, besides the prescription of long-acting benzodiazepines, also the prescription of intermediate or short-acting benzodiazepines, which can cause adverse reactions such as falls, fractures, and cognitive deterioration in elderly patients³⁸. The second cause of PIP, according to Beers criteria ID, was the use of non-COX-2 selective non-steroidal anti-inflammatory drugs, which coincides with other studies^28,31. The prolonged use of non-steroidal anti-inflammatory drugs forchronic mild pain in osteoarthritis was also present to ahigh rate in the STOPP Criteria, coinciding with previousstudies^24,39. Besides, acetylsalicylic acid was involved asthe third most frequent STOPP criteria due to excessiveuse, thus confirming the variability in antiaggreganttreatments for primary and secondary cardiovascularprevention.

Collecting the prescription invoicing data confirmed through OMI-AP electronic clinical records has allowed us to conduct an in-depth study of PIP frequency, and to offer reliable information about patients’ chronic treatment. However, our study presents the limitation of the insufficient completion of some clinical records, which made it difficult to identify some criteria. Another limitation was the lack of real assessment of the clinical signs of patients, in order to evaluate, for example, the presence of oedemas or the intensity of articular pain in osteoarthritis.

In 2015, while this study was being conducted, new updates were published for Beers criteria17 and STOPP/ START criteria¹⁹. If we had applied these new versions, our results might have been different, maybe with an even higher PIP rate detected through both criteria.

In conclusion, we have observed there is a high prevalence of PIP in highly polypharmacyolder patients, detected both with Beers criteria and STOPP criteria, The most frequent causes for PIP were therapeutic duplications, as well as the conditions of use for benzodiazepines, an- tiaggregant drugs, and non-steroidal anti-inflammatory drugs. In our opinion, the simultaneous application of Beers and STOPP criteria represents a useful tool for PIP screening in polymedicated community-dwelling older patients.

Contribution to Scientific Literature: Polypharmacy patients represent a population group prone to receiving potentially inappropriate prescriptions. Beers and STOPP Criteria are explicit methods to assess the suitability of pharmacological treatment. This study shows a frequent prescription of therapeutic duplication, benzodiazepines, and acetylsalicylic acid, without clear indication, as well as non-steroidal anti-inflammatory drugs, in polypharmacy> 65-year-old patients.

How to cite this article: Nicieza-García ML, Salgueiro-Vázquez ME, Jimeno-Demuth FJ, Manso G.Beers versus STOPP criteria in polyharmacy community-dwelling older patients.Farm Hosp. 2016;40(3):150-164.