Prevalence of Potentially Inappropriate Medications in patients with multimorbidity according to LESS-CHRON and STOPPFrail criteria

Journal of the American Medical Directors Association Prevalence of PIM in patients with multimorbidity according to LESS-CHRON and STOPPFrail criteria --Manuscript Draft-- Manuscript Number: JAMDA-D-22-00659R2 Article Type: Original Study - Brief Report Keywords: LESS-CHRON; STOPPFrail; Deprescribing; Older patients; multimorbidity; polymedicated. Corresponding Author: Aitana Rodriguez-Perez, PhD Virgen del Rocio University Hospital SPAIN First Author: Marta Mejías-Trueba Order of Authors: Marta Mejías-Trueba Aitana Rodriguez-Perez, PhD Irene Sotillo-Sánchez Susana Sánchez-Fidalgo Maria Dolores Nieto-Martín Emilio García-Cabrera Manuscript Region of Origin: Abstract: Abstract Objective: LESS-CHRON and STOPPFrail are criterion-based deprescribing tools. This study aimed to identify the prevalence of potentially inappropriate medications (PIM) with these tools in an outpatient, polymedicated, older population with multimorbidity. Design: Single-center cross-sectional observational study. Setting and Participants: PIM and criteria subject to deprescribing identified by each tool were collected in patients who were being followed up on outpatient internal medicine consultation. Methods: PIM were identified by STOPPFrail and LESS-CHRON criteria reviewing medical histories and pharmacological treatments of the patients in the electronic health card system. Sociodemographic, clinical and pharmacological variables were recorded. A correlation analysis between treatment tools and clinical values was performed using the non-parametric Spearman’s Rho correlation. Results: Eighty-three patients with a median of 14.4 (12;17) prescribed drugs were included. The total number of PIM identified with LESS-CHRON was 158 vs. 127 with STOPPFrail. Eight of the 27 criteria (29.6%) for LESS-CHRON and 15 of the 25 for STOPPFrail were found to be not applicable. A significant correlation was obtained for both tools with the number of prescribed drugs at the time of inclusion. The Profund, Barthel, and VGI-index only showed a significant correlation with LESS-CHRON. Conclusion and Implications: Both tools have shown the capacity to identify PIM that can be deprescribed in the population studied. However, LESS-CHRON appears to have a greater detection potential in the subgroup of patients analyzed. STOPPFrail brings a certain complementarity in other areas of therapy not covered by LESSCHRON. Powered by Editorial Manager® and ProduXion Manager® from Aries Systems Corporation 08/08/2022 I am writing as author of correspondence of the manuscript titled “Prevalence of PIM in patients with multimorbidity according to LESS-CHRON and STOPPFrail criteria.”. This manuscript compares the detection potential of the STOPPFrail criteria versus LESS-CHRON as screening tools in a cohort of polymedicated elderly patients with multimorbidity. We analyse the ability of these tools to identify inappropriate drugs in this population, as well as the therapeutic groups most frequently involved in the deprescription opportunities detected. In addition, the evaluation of the correlation between the different deprescription criteria identified, the deprescription opportunities and the applied clinical scales was carried out. This work highlights the theoretical usefulness of two deprescription tools that are very well positioned in the scientific literature. This work is not under active consideration for publication, has not been accepted for publication, nor has it been published, in full or in part (except in abstract form). I confirm that the study has been approved by the Ethics Committee of Virgen del Rocío University Hospital, an institutional ethics committee For all this we remain attentive to your response. If you need any further information, do not hesitate to contact me. We explicitly state that: (1) the manuscript has not been and will not be submitted, in part or entirety, elsewhere for publication; (2) all authors meet criteria for authorship as stated in the Uniform Requirements for Manuscripts Submitted to Biomedical Journals (explained below), as well as their contributions to the manuscript; (3) if accepted, the paper will not be published elsewhere in the same form, in English or in any other language, including electronically, without the written consent of the copyright holder; (4) the instructions for authors have been taken into account, including that all signatory authors meet the requirements of authorship and declare that they have no conflict of interest. In addition, a list of three potential reviewers for our manuscript are described: - Srtefan Zechmann: email [email protected], Institute of Primary Care, University of Zurich, Pestalozzistrasse 24, 8091, Zurich, Switzerland Cover Letter Click here to access/download;Cover Letter;Cover letter.docx - Y.S. Karandikar: [email protected], Department of Pharmacology, B.V.D.U. Medical College, Pune, India - J González-Bueno: email: [email protected], Department of Pharmacy, Hospital Universitari de Vic (Consorci Hospitalari de Vic), Vic (Barcelona), Spain. Regards, Aitana Rodríguez Pérez (corresponding author) PhD Clinical Pharmacist Tel.: (+34) 955013622 [email protected] UGC Farmacia. Hospital Universitario Virgen del Rocío Avda Manuel Siurot s/n. 41013 Sevilla (SPAIN) Pharmacy Service Virgen del Rocío Hospital Seville Spain, 7th November 2022 Dear reviewers: We deeply appreciated the exhaustive review and the contributions raised in our work presented in the manuscript Prevalence of PIM in patients with multimorbidity according to LESS-CHRON and STOPPFrail criteria (JAMDA-D-22-00659). The corrected version of the manuscript with the modifications, made and indicated with “track changes” has been send. Below are our comments for the reviewer. We hope we have addressed all issues raised by the reviewer. However, do not hesitate to contact us if the need further information. REVIEW 2: 1. Reviewer #2: Thanks to the authors for the revised manuscript. The responses that describe how STOPPFrail and LESS-CHRON are implemented and how to interpret the figures are particularly helpful, and I would recommend adding a supplemental appendix to describe the circuit (although I defer to the editor—perhaps this isn't necessary for the JAMDA audience). Following your recommendations, we have included the circuit carried out as supplementary material (S3). If the editor considers it necessary to include also the interpretation of the figure, we add it without problems. 2. The organization of the Discussion section is helpful as well, which I think would benefit from a discussion of the fixed costs of the two platforms so that readers could have more context for deciding which one to implement. Thank you very much, certainly the changes in the discussion have improved the article. Regarding the issue of costs, the truth is that, to the best of our knowledge, there is no web platform available for the STOPPFrail criteria, just the original article with the criteria is published. In the case of LESS-CHRON, it is available in web format (https://chronicpharma.com/) and it is free. We have not added anything in this regard because the authors consider that both are easy to apply since they are available (supplementary material S1 and S2 of this manuscript). Adding this information perhaps benefited LESS-CHRON excessively and did not seem entirely fair to us. Despite this, we have added a sentence to the discussion in this regard: First of all, it is important to mention that both tools are available in the scientific literature without any associated costs. Response to reviewer 3. I am still concerned with selection bias and the possibility that the study sample mechanically increase the likelihood that LESS-CHRON dominates in the head-to- head. However, I recognize that the authors are limited in space and may not be able to conduct more rigorous analysis given the setting and statistical power. I was confused by the authors' statement in the response document that their sample is representative, which lacked evidence, and their claim that they are "avoiding selection bias," which is impossible. We are sorry for the misunderstanding, perhaps we did not express ourselves correctly in the previous letter. It is true that if we do not randomize, bias is more likely to occur. However, our sample is representative of chronic complex patients because the entire sample has these characteristics, and the appropriate sample size is reached according to the calculation performed: “To achieve a correlation coefficient other than 0 and an r of at least 0.30, assuming a type I error of 0.05 and a type II error of 0.20, it was necessary to include at least 80 subjects in the study”. In other words, we tried to indicate that the inclusion of 83 patients represents a sufficient sample in terms of statistical power. In addition, the main objective of this study was to compare the potential theoretical applicability of these tools in a population on which we can routinely make pharmaceutical interventions. Therefore, while we understand your concern, we believe that we have not committed any significant bias. As we mentioned earlier, when we reviewed candidate patients, we included all those who had any potential PIM for any of the tools evaluated. Therefore, there are patients who only have applicable PIM for STOPPFrail and not for LESSCHRON. Thus, when applying theoretically and simultaneously both tools to each patient, if STOPPFrail would have had a very different potency than LESS-CHRON, this would be reflected in these results. 4. The authors then state that "both tools are perfectly comparable in terms of methodology and design" but describe reasons why they are not comparable. Indeed, perhaps the statement is contradictory, we explain. We are referring to the fact that, as has been demonstrated in previously published works, both tools have been shown to be comparable in terms of methodology and design (10.1016/j.sapharm.2022.03.008; 10.1111/jgs.15616). That is, both are explicit criteria, focused on elderly patients, with the objective of deprescribing drugs (not focused on appropriateness), designed by a panel of experts following a Delphi methodology, organized by therapeutic groups, etc. However, this does not exempt that there are other aspects that slightly differentiate them, such as the fact that STOPPFrail includes two not explicit criteria (section A), which were eliminated. Nevertheless, as we mentioned in the previous letter, it is common to find articles in the literature comparing tools that are somewhat different from each other (LESS-PHARMA Study), which does not mean that it cannot be done as long as it is properly discussed as we have done. However, our objective is to compare the detection potential of both tools and to describe the main differences and similarities, which may logically influence the applicability of each one, but hence the value we intend to contribute with this work. 5. It also appeared that the authors de-emphasize that STOPPFrail is specifically geared toward frail adults in the introduction. Totally, we have deleted it from the previous version in an attempt to simplify. We have added it again: On the other hand, STOPPFrail is specifically geared toward older frail patients. This tool also takes into account clinical conditions to perform deprescribing, although it does not specify out health variables or monitoring parameters15. Prevalence of PIM in patients with multimorbidity according to LESS-CHRON and STOPPFrail criteria Marta Mejías-Trueba PhD1, Aitana Rodríguez-Pérez PhD*1, Irene Sotillo-Sánchez PharmD2, Susana Sánchez-Fidalgo PhD2, Maria Dolores Nieto- Martin MD, PhD3, Emilio García-Cabrera PhD2 1 Pharmacy Clinical Management Unit. Virgen del Rocío University Hospital. Sevilla, Spain 2 Department of Preventive Medicine and Public Health, University of Seville. Sevilla, Spain. 3 Internal Medicine Clinical Management Unit. Virgen del Rocío University Hospital. Sevilla, Spain Correspondence: aitan[email protected]; Avda Manuel Siurot s/n. 41013-Sevilla (Spain) Tel.: (0034) 95502095. Running title: Prevalence of PIM according to deprescribing tools Key words: LESS-CHRON; STOPPFrail; deprescribing; older patients; multimorbidity; polymedicated. Funding sources: MMT received financial support from the Subprograma Río Hortega, Instituto de Salud Carlos III, Subdirección General de Redes y Centros de Investigación Cooperativa, Ministerio de Ciencia, Innovación y Universidades, Spain (CM21/00115) Word, reference, and graphics count: 2630 words, 22 references, 3 tables and 1 figure. Title Page (Containing author details) Brief summary: Inappropriate medication is highly prevalent among elderly and chronic patients. Deprescription tools have been development as strategies to identify it. Studies detecting potentially inappropriate medications are necessary. Acknowledgements: The authors thank Máximo Bernabeu-Wittel, Manuel Ollero- Baturone, Carlos Hernández-Quiles, Bosco Barón-Franco and Bernardo Santos-Ramos for clinical help. Prevalence of PIM in patients with multimorbidity according to LESS-CHRON and 1 STOPPFrail criteria 2 Abstract 3 Objective: LESS-CHRON and STOPPFrail are criterion-based deprescribing tools. This 4 study aimed to identify the prevalence of potentially inappropriate medications (PIM) 5 with these tools in an outpatient, polymedicated, older population with multimorbidity. 6 Design: Single-center cross-sectional observational study. 7 Setting and Participants: PIM and criteria subject to deprescribing identified by each 8 tool were collected in patients who were being followed up on outpatient internal 9 medicine consultation. 10 Methods: PIM were identified by STOPPFrail and LESS-CHRON criteria reviewing 11 medical histories and pharmacological treatments of the patients in the electronic health 12 card system. Sociodemographic, clinical and pharmacological variables were recorded. 13 A correlation analysis between treatment tools and clinical values was performed using 14 the non-parametric Spearman’s Rho correlation. 15 Results: Eighty-three patients with a median of 14.4 (12;17) prescribed drugs were 16 included. The total number of PIM identified with LESS-CHRON was 158 vs. 127 with 17 STOPPFrail. Eight of the 27 criteria (29.6%) for LESS-CHRON and 15 of the 25 for 18 STOPPFrail were found to be not applicable. A significant correlation was obtained for 19 both tools with the number of prescribed drugs at the time of inclusion. The Profund, 20 Barthel, and VGI-index only showed a significant correlation with LESS-CHRON. 21 Conclusion and Implications: Both tools have shown the capacity to identify PIM that 22 can be deprescribed in the population studied. However, LESS-CHRON appears to have 23 Manuscript (clean) Click here to view linked References and ischemic heart disease (30.1%), respectively. With regard to the comorbidities 154 recorded, the most common were arterial hypertension (88%), dyslipidemia (60.2%), 155 atrial fibrillation (48.2%), and diabetes mellitus without complications (45.8%). 156 Assessment deprescribing with the LESS-CHRON and STOPPFrail tools 157 Regarding pharmacological treatment, as has already been commented, the median 158 number of prescribed drugs per patient was 14.4 (12;17). The total number of PIM and 159 deprescribing criteria identified with both tools are shown in Table 2. No significant 160 differences were detected between the tools. 161 Of the cohort studied, 10.8% of patients (n=9) did not meet any deprescribing criteria for 162 the LESS-CHRON tool and 13.3% (n=11) did not for the STOPPFrail tool. 163 The prevalence of PIM according to deprescribing criteria of each tool is presented in 164 Figure 1. 165 In LESS-CHRON, 8 of the 27 criteria (29.6%) were found not to be applicable to any 166 patient. The most prevalent criterion was withdrawal from benzodiazepine for insomnia 167 (F3), with a value of 31%, followed closely by the deprescribing of antidepressants (F4) 168 in the case of reactive depression (27%). In the case of STOPPFrail, 15 of the 25 criteria 169 (60%) were not used and the withdrawal criterion of the lipid-lowering therapy (B1) was 170 the most theoretically applied, present in 69% of patients. 171 Correlation analysis between pharmacological and clinical variables in both tools 172 A weak correlation between both tools was observed (29.5%, p=0.007 with respect to the 173 criteria; 29%, p=0.008 with respect to the PIM for deprescribing) (Table 3A). As far as 174 the correlation between the prescribed drugs in the population at the time of inclusion, 175 although there is a correlation in both tools, this was found to be greater in LESS- 176 CHRON. 177 Table 3B shows that none of the clinical indexes correlated with the STOPPFrail tool. 178 However, the Profund, Barthel, and frailty indexes did show a significant correlation with 179 LESS-CHRON for both the criteria variable and PIM for deprescribing. 180 Discussion 181 Population target 182 This study, which is the first to compare LESS-CHRON and STOPPFrail tools, confirms 183 that both explicit tools can be used to detect PIM in frail patients or those with a limited 184 life expectancy13. This is based on the fact that 77% of patients (n=64) met a minimum 185 of one deprescribing criterion in both tools. These results are consistent with the literature 186 and show the usefulness of deprescription tools in supporting the process16, 22. However, 187 the LESS-CHRON tool was found to have greater detection potential due to a higher 188 number of PIM obtained, and the correlation of the tool with pharmacological and clinical 189 parameters. 190 The differences found could be due to the target population towards which both tools are 191 directed. In this sense, the STOPPFrail authors geared the tool towards frail patients with 192 a life expectancy less than one year15. However, LESS-CHRON was designed to be 193 applied in chronic patients with multimorbidity or those with complex health needs14, 194 which does not necessarily imply a short survival prognosis. In fact, survival prognosis is 195 included in this tool as a variable to consider in different proposed deprescribing criteria. 196 The cohort analyzed in this study is older, with a high degree of polymedication and 197 comorbidities. However, in terms of prognosis, the Profund index obtained implies a low- 198 intermediate risk of mortality at 12 months and Frail-VIG index indicates that elevated 199 mortality at 1 year cannot be presumed19, 20. 200 Theoretical applicability of the tools 201 Taking into account the theoretical applicability of the tools, combined to the fact that 202 STOPPFrail15 incorporates laxer withdrawal criteria, it could be expected that the number 203 of PIM detected with this tool is greater than with LESS-CHRON. However, the results 204 obtained show the opposite. This could be the consequence of a greater suitability of the 205 LESS-CHRON criteria for this population or the tool having a greater sensitivity for the 206 type of the prescribed medication in these patients. In this way, although the criteria of 207 the STOPPFrail tool are more flexible, if the patient is not clearly at the end of their life, 208 the application may be somewhat complex. This, together with the elimination of section 209 A, could determine the lower detection capacity of PIM compared to previous studies of 210 this tool23. 211 Differences according to the functional system 212 In line with the criteria for each tool according to the drug involved, the differences 213 according to the functional system can be analyzed. Consequently, the most notable is 214 found in the systems or groups of drugs that are only represented in one of the tools. 215 Specifically, in the case of STOPPFrail, the criteria related to the gastrointestinal tract 216 (PPI and H2 antagonists), which make up 21% of the total, and in the case of LESS- 217 CHRON, the group of benzodiazepines, Z drugs, and antidepressants (46% of the total), 218 together with the scenario that suggests the deprescribing of allopurinol in the secondary 219 prevention of gout (11%). In fact, they can be considered the criteria with greatest 220 complementarity between both tools, representing a high percentage of the total criteria 221 identified (40.8%). 222 Similarities between LESS-CHRON and STOPPFrail 223 First of all, it is important to mention that both tools are available in the scientific literature 224 without any associated costs. 225 Regarding the similarities identified, it is striking that they share 10 of the same criteria, 226 although they slightly differ with respect to the clinical situation that allows for 227 deprescribing. If the criterion referring to lipid-lowering therapies is specifically 228 analyzed, it is the most applied in STOPPFrail (46% of the total), this figure is not 229 maintained for LESS-CHRON since this tool establishes more restrictive conditions. The 230 same is true for the third most prevalent criterion in STOPPFrail, where general 231 withdrawal of calcium supplements is advocated, while LESS-CHRON specifies that 232 patients must have a Barthel index <60 and be unable to walk, again creating differences 233 in applicability (20% vs. 7%). In the opposite situation, the criterion indicating the 234 deprescribing of alpha-blockers in prostatic hypertrophy is highly prevalent in LESS- 235 CHRON, while in STOPPFrail, withdrawal was only recommended when the patient was 236 catheterized. Finally, in the endocrine system, both tools indicate the withdrawal of oral 237 antidiabetics, but due to differences in clinical conditions, the prevalence in LESS- 238 CHRON is 14% vs. 8% in STOPPFrail. 239 Another relevant finding is that although the number of criteria analyzed for each tool 240 was similar (27 in LESS-CHRON and 25 in STOPPFrail), it is true that in STOPPFrail 241 only two criteria made up 65% of the total (criteria B1 and E1) and one of them affected 242 70% of the patients (criterion B1). However, in LESS-CHRON, the application of the 243 different items was more homogeneous, distributed over a greater number of different 244 criteria, and the sum of the three most applied criteria was less than 50% of the total. 245 It should be note that while this study was conducted, a new version of the STOPPFrail 246 tool was published24, which incorporates new criteria and eliminates others that had 247 become obsolete. Nevertheless, none of the previously mentioned criteria that carried 248 considerable weight in LESS-CHRON have been added in the new version. 249 Finally, it is important to mention that both tools consider criteria did not detect any 250 candidate drug to be withdrawn, with a higher number in STOPPFrail compared to LESS- 251 CHRON (15 vs 8 criteria). However, after the update to the English version of 252 STOPPFrail17, 6 of those 15 criteria have been eliminated, while 2 of them have been 253 modified. With respect to LESS-CHRON, a study is currently underway to validate the 254 criteria in a cohort of institutionalized patients and outpatients. With this study, data on 255 usability and clinical utility will be obtained that will allow the criteria to be adapted to 256 actual use needs. 257 In assessing the correlation between the clinical parameters, it is worth noting that neither 258 the number of PIM nor the STOPPFrail criteria showed any sort of correlation with frailty, 259 while there was a weak and significant correlation with the LESS-CHRON criteria 260 (28.6%, p=0.014). This may be due to the fact that the degree of frailty of the patient in 261 our cohort did not exceed the threshold for which the STOPPFrail tool was designed. 262 At the same time, prognostic assessment19 and Barthel index18 again showed a weak but 263 significant correlation with the LESS-CHRON criteria, but no correlation with the 264 STOPPFrail tool. This could be explained by the fact that the LESS-CHRON criteria 265 incorporate as part of the clinical evaluation of patients. 266 Consequently, one can deduce that the LESS-CHRON tool better adapts to the patient’s 267 clinical and prognostic context, including a greater number of criteria in older chronic 268 patients with multimorbidity. However, given the low correlation rates, it would be 269 desirable to design new studies with a larger sample focused on this aspect. 270 Limitations 271 Among the limitations of the study, we highlight that the most recent version of 272 STOPPFrail was not used. In addition, section A was removed from it in order to make a 273 more realistic comparison between deprescribing criteria. Expanding this study to a larger 274 number of patients would provide more reliable data on the correlation between the tools 275 and the clinical parameters analyzed. 276 Conclusions and Implications 277 In summary, both tools have demonstrated a great capacity to identify PIM that can be 278 deprescribed in an older polymedicated population with multimorbidity. In addition, they 279 complement each other, covering a greater number of areas of pharmacotherapy between 280 the two. Though, because STOPPFrail was designed and validated in a population with a 281 poor prognosis and great frailty, LESS-CHRON appears to have greater detection 282 potential in the subgroup of patients analyzed. It is necessary to confirm these results in 283 the context of applying a real-life criteria application. 284 Conflicts of Interest 285 The authors declare that there is no conflict of interest regarding the publication of this 286 article. 287 References 288 1. Zemedikun DT, Gray LJ, Khunti K, Davies MJ, Dhalwani NN. Patterns of 289 Multimorbidity in Middle-Aged and Older Adults: An Analysis of the UK Biobank Data. 290 Mayo Clin Proc. 2018;93(7):857-866. doi: 10.1016/j.mayocp.2018.02.012. 291 2. Chudasama YV, Khunti K, Gillies CL, Dhalwani NN, Davies MJ, Yates T, et al. 292 Healthy lifestyle and life expectancy in people with multimorbidity in the UK Biobank: 293 A longitudinal cohort study. PLoS Med. 2020;17(9):e1003332. https:// 294 doi.org/10.1371/journal.pmed.1003332. 295 3. Zazzara MB, Vetrano DL, Carfì A, Onder G. Frailty and chronic disease. Panminerva 296 Med. 2019;61(4):486-492. doi: 10.23736/S0031-0808.19.03731-5. 297 4. Hanlon P, Nicholl BI, Jani BD, Lee D, McQueenie R, Mair FS. Frailty and pre-frailty 298 in middle-aged and older adults and its association with multimorbidity and mortality: a 299 prospective analysis of 493 737 UK Biobank participants. Lancet. 2018;3(7):323–332. 300 doi: 10.1016/S2468-2667(18)30091-4. 301 5. Barnett K, Mercer SW, Norbury M, Watt G, Wyke S, Guthrie B. Epidemiology of 302 multimorbidity and implications for health care, research, and medical education: a cross- 303 sectional study. Lancet. 2012;380: 37–43. doi: 10.1016/S0140-6736(12)60240-2. 304 6. Bernabeu-Wittel M, Ollero-Baturone M, Nieto-Martín D, García-Morillo S, 305 Goicoechea-Salazar J. Patient-Centered Care for Older Adults with Multiple Chronic 306 Conditions: These are the Polypathological Patients! J Am Geriatr Soc. 2013;61(3):475- 307 6. doi: 10.1111/jgs.12142. 308 7. Bernabeu Wittel M, Barón Franco B, Murcia Zaragoza J, Fuertes Martín A, Ramos 309 Cantos C, Fernández Moyano A, et al. A multi-institutional, hospital-based assessment 310 of clinical, functional, sociofamilial and health-care characteristics of polypathological 311 patients (PP). Arch Gerontol Geriatr. 2011;53(3):284–91. doi: 312 10.1016/j.archger.2010.12.006. 313 8. Page AT, Falster MO, Litchfield M, Pearson SA, Etherton-Beer CH. Polypharmacy 314 among older Australians, 2006-2017: a population-based study. Med J Aust. 315 2019;211(2):71-75. doi: 10.5694/mja2.50244 316 9. Martínez-Sotelo J, Pinteño-Blanco M, García-Ramos R, Llobera-Cànaves J, Cadavid- 317 Torres MI. Pharmacist-led intervention on potentially inappropriate prescription in 318 patients with polypharmacy: PHARM-PC clinical trial protocol. Farm Hosp. 319 2021;45(4):210-215. English. doi: 10.7399/fh.11575. 320 10. Gallagher J, O’Sullivan D, McCarthy S, Gillespie P, Woods N, O’Mahony D, et al. 321 Structured pharmacist review of medication in older hospitalised patients: a cost- 322 effectiveness analysis. Drugs Aging. 2016; 33(4): 285–294. 323 11. Ulley J, Harrop D, Ali A, Alton S, Fowler Davis S. Deprescribing interventions and 324 their impact on medication adherence in community-dwelling older adults with 325 polypharmacy: a systematic review. BMC Geriatr. 2019;18;19(1):15. doi: 326 10.1186/s12877-019-1031-4. 327 12. Rodríguez Pérez A, Alfaro Lara ER, Nieto Martín MD, Ruiz Cantero A, Santos 328 Ramos B. Deprescribing in patients with multimorbidity: A necessary process. European 329 Journal of Internal Medicine. 2015;26(7):e18-e19. doi: 10.1016/j.ejim.2015.06.011. 330 13. Thompson W, Lundby C, Graabæk T, Nielsen DS, Ryg J, Søndergaard J et al. Tools 331 for Deprescribing in Frail Older Persons and Those with Limited Life Expectancy: A 332 Systematic Review. J Am Geriatr Soc. 2019;67(1):172-180. doi: 10.1111/jgs.15616. 333 14. Rodríguez Pérez A, Alfaro Lara ER, Albiñana Pérez S, Nieto Martín MD, Díez 334 Manglano J, Pérez Guerrero C et al. Novel tool for deprescribing in chronic patients with 335 multimorbidity: List of Evidence-Based Deprescribing for Chronic Patients criteria. 336 Geriatr Gerontol Int. 2017;17(11):2200-2207. doi: 10.1111/ggi.13062. 337 15. Lavan A, Gallagher P, Parsons C et al. STOPPFrail (Screening Tool of Older Persons 338 Prescriptions in Frail adults with limited life expectancy): Consensus validation. Age 339 Ageing 2016;45:1–12. DOI: 10.1093/ageing/afx005 340 16. Curtin D, Dukelow T, James K, O'Donnell D, O'Mahony D, Gallagher P. 341 Deprescribing in multi-morbid older people with polypharmacy: agreement between 342 STOPPFrail explicit criteria and gold standard deprescribing using 100 standardized 343 clinical cases. Eur J Clin Pharmacol. 2019 Mar;75(3):427-432. doi: 10.1007/s00228-018- 344 2598-y. 345 17. Pfeiffer E. A short portable mental status questionaire for the assesment of organic 346 brain deficit in elderly patients. J Am Geriatr Soc 1975; 23: 433-441. 347 18. Mahoney FI, Barthel DW. Functional evaluation: the Barthel Index. Md State Med J 348 1965;4:61–5. 349 19. Bernabeu Wittel M, Moreno Gaviño L, Ollero Baturone M, Barón Franco B, Díez 350 Manglano J, Rivas-Cobas C, et al. Validation of PROFUND prognostic index over a 351 fouryear follow-up period. Eur J Intern Med. 2016;36:20–4. 352 20. Amblàs Novellasa J, Martori JC, Molist Brunet N, Oller R, Gómez-Batiste X, 353 Espaulella Panicot J. [Frail-VIG index: Design and evaluation of a new frailty index based 354 on the Comprehensive Geriatric Assessment].Rev Esp Geriatr Gerontol. 2017;52(3):119- 355 127. doi: 10.1016/j.regg.2016.09.003. 356 21. Hulley SB, Cummings SR, Browner WS, Grady D, Newman TB. Designing clinical 357 research: an epidemiologic approach. 4th ed. Philadelphia, PA: Lippincott Williams & 358 Wilkins. 2013;6C:79. 359 22. Rodríguez-Pérez A, Alfaro-Lara ER, Sierra-Torres MI, Villalba-Moreno Á, Nieto- 360 Martin MD, Galván-Banqueri M, Santos-Ramos B. Validation of the LESS-CHRON 361 criteria: reliability study of a tool for deprescribing in patients with multimorbidity. Eur 362 J Hosp Pharm. 2019 Nov;26(6):334-338. doi: 10.1136/ejhpharm-2017-001476. 363 23. Hyun-Woo Ch, Yoonhee K, Yewon S, Junghwa L, Eunsook L, Euni L, et al. 364 Prevalence of potentially inappropriate medications based on the STOPPFrail criteria in 365 frail older patients with limited life expectancy: a cross-sectional study. BMC Geriatr. 366 2022 Apr 27;22(1):367. doi: 10.1186/s12877-022-03067-7. 367 24. Curtin D, Gallagher P, O'Mahony D. Deprescribing in older people approaching end- 368 of-life: development and validation of STOPPFrail version 2. Age Ageing. 369 2021;50(2):465-471. doi: 10.1093/ageing/afaa159. 370 371 372 373 374 Figure 1. A) Prevalence* of PIM with the STOPPFrail tool. 375 B) Prevalence* of PIM with the LESS-CHRON tool. The physiological systems to which 376 the criteria belonged were identified with an alphanumeric code (Supplementary Material 377 S1 and S2). Categories A1 and A2 of STOPPFrail were not assessed because they are 378 appropriateness criteria. 379 *Prevalence (%) was calculated as the ratio between the number of potentially 380 inappropriate medications (PIM) identified for each criterion versus the total number of 381 individuals included (n=83). 382 To achieve a correlation coefficient other than 0 and an r of at least 0.30, assuming a type 132 I error of 0.05 and a type II error of 0.20, it was necessary to include at least 80 subjects 133 in the study21. 134 Ethics approval 135 The study was conducted in accordance with the Declaration of Helsinki and approved 136 by the Ethics Committee of XXX. Informed consent was obtained from all subjects 137 involved in the study. 138 Results 139 Of all the patients who agreed to participate in the study, 83 patients met all inclusion 140 criteria and then were included in the study. However, certain clinical variables were not 141 recorded in some patients as reflected in Table I. 142 Demographic, pharmacological, and clinical characteristics 143 Of the patients analyzed, 81.5% did not show cognitive decline according to the Pfeiffer 144 questionnaire. Similarly, the median of Barthel index showed slight dependence, ranging 145 from 60 to 95 points, in 52.4% of the patients. In addition, the median obtained with the 146 Profund index suggested a low-intermediate risk (21.5-31.5%) of death at one year after 147 hospital discharge in 33.7% of the patients and an intermediate-high risk (45-50%) in 148 25.3%. The patient cohort showed mild frailty, but near the upper limit (mild frailty 0.2- 149 0.36 and moderate frailty 0.39-0.53). 150 Based on established inclusion criteria, the population studied had a high prevalence of 151 chronic disease. Thus, the median number of clinical categories presented was 3 (3;4). 152 The most frequent clinical categories were heart failure (63.9%), chronic kidney disease 153 (55.4%), anemia from gastrointestinal loss (48.2%), chronic respiratory disease (39.8%), 154 and ischemic heart disease (30.1%), respectively. With regard to the comorbidities 155 recorded, the most common were arterial hypertension (88%), dyslipidemia (60.2%), 156 atrial fibrillation (48.2%), and diabetes mellitus without complications (45.8%). 157 Assessment deprescribing with the LESS-CHRON and STOPPFrail tools 158 Regarding pharmacological treatment, as has already been commented, the median 159 number of prescribed drugs per patient was 14.4 (12;17). The total number of PIM and 160 deprescribing criteria identified with both tools are shown in Table 2. No significant 161 differences were detected between the tools. 162 Of the cohort studied, 10.8% of patients (n=9) did not meet any deprescribing criteria for 163 the LESS-CHRON tool and 13.3% (n=11) did not for the STOPPFrail tool. 164 The prevalence of PIM according to deprescribing criteria of each tool is presented in 165 Figure 1. 166 In LESS-CHRON, 8 of the 27 criteria (29.6%) were found not to be applicable to any 167 patient. The most prevalent criterion was withdrawal from benzodiazepine for insomnia 168 (F3), with a value of 31%, followed closely by the deprescribing of antidepressants (F4) 169 in the case of reactive depression (27%). In the case of STOPPFrail, 15 of the 25 criteria 170 (60%) were not used and the withdrawal criterion of the lipid-lowering therapy (B1) was 171 the most theoretically applied, present in 69% of patients. 172 Correlation analysis between pharmacological and clinical variables in both tools 173 A weak correlation between both tools was observed (29.5%, p=0.007 with respect to the 174 criteria; 29%, p=0.008 with respect to the PIM for deprescribing) (Table 3A). As far as 175 the correlation between the prescribed drugs in the population at the time of inclusion, 176 although there is a correlation in both tools, this was found to be greater in LESS- 177 CHRON. 178 Table 3B shows that none of the clinical indexes correlated with the STOPPFrail tool. 179 However, the Profund, Barthel, and frailty indexes did show a significant correlation with 180 LESS-CHRON for both the criteria variable and PIM for deprescribing. 181 Discussion 182 Population target 183 This study, which is the first to compare LESS-CHRON and STOPPFrail tools, confirms 184 that both explicit tools can be used to detect PIM in frail patients or those with a limited 185 life expectancy13. This is based on the fact that 77% of patients (n=64) met a minimum 186 of one deprescribing criterion in both tools. These results are consistent with the literature 187 and show the usefulness of deprescription tools in supporting the process16, 22. However, 188 the LESS-CHRON tool was found to have greater detection potential due to a higher 189 number of PIM obtained, and the correlation of the tool with pharmacological and clinical 190 parameters. 191 The differences found could be due to the target population towards which both tools are 192 directed. In this sense, the STOPPFrail authors geared the tool towards frail patients with 193 a life expectancy less than one year15. However, LESS-CHRON was designed to be 194 applied in chronic patients with multimorbidity or those with complex health needs14, 195 which does not necessarily imply a short survival prognosis. In fact, survival prognosis is 196 included in this tool as a variable to consider in different proposed deprescribing criteria. 197 The cohort analyzed in this study is older, with a high degree of polymedication and 198 comorbidities. However, in terms of prognosis, the Profund index obtained implies a low- 199 intermediate risk of mortality at 12 months and Frail-VIG index indicates that elevated 200 mortality at 1 year cannot be presumed19, 20. 201 Theoretical applicability of the tools 202 Taking into account the theoretical applicability of the tools, combined to the fact that 203 STOPPFrail15 incorporates laxer withdrawal criteria, it could be expected that the number 204 of PIM detected with this tool is greater than with LESS-CHRON. However, the results 205 obtained show the opposite. This could be the consequence of a greater suitability of the 206 LESS-CHRON criteria for this population or the tool having a greater sensitivity for the 207 type of the prescribed medication in these patients. In this way, although the criteria of 208 the STOPPFrail tool are more flexible, if the patient is not clearly at the end of their life, 209 the application may be somewhat complex. This, together with the elimination of section 210 A, could determine the lower detection capacity of PIM compared to previous studies of 211 this tool23. 212 Differences according to the functional system 213 In line with the criteria for each tool according to the drug involved, the differences 214 according to the functional system can be analyzed. Consequently, the most notable is 215 found in the systems or groups of drugs that are only represented in one of the tools. 216 Specifically, in the case of STOPPFrail, the criteria related to the gastrointestinal tract 217 (PPI and H2 antagonists), which make up 21% of the total, and in the case of LESS- 218 CHRON, the group of benzodiazepines, Z drugs, and antidepressants (46% of the total), 219 together with the scenario that suggests the deprescribing of allopurinol in the secondary 220 prevention of gout (11%). In fact, they can be considered the criteria with greatest 221 complementarity between both tools, representing a high percentage of the total criteria 222 identified (40.8%). 223 Similarities between LESS-CHRON and STOPPFrail 224 First of all, it is important to mention that both tools are available in the scientific literature 225 without any associated costs. 226 Regarding the similarities identified between both tools, it is striking that they share 10 227 of the same criteria, although they slightly differ with respect to the clinical situation that 228 allows for deprescribing. If the criterion referring to lipid-lowering therapies is 229 specifically analyzed, it is the most applied in STOPPFrail (46% of the total), this figure 230 is not maintained for LESS-CHRON since this tool establishes more restrictive 231 conditions. The same is true for the third most prevalent criterion in STOPPFrail, where 232 general withdrawal of calcium supplements is advocated, while LESS-CHRON specifies 233 that patients must have a Barthel index <60 and be unable to walk, again creating 234 differences in applicability (20% vs. 7%). In the opposite situation, the criterion 235 indicating the deprescribing of alpha-blockers in prostatic hypertrophy is highly prevalent 236 in LESS-CHRON, while in STOPPFrail, withdrawal was only recommended when the 237 patient was catheterized. Finally, in the endocrine system, both tools indicate the 238 withdrawal of oral antidiabetics, but due to differences in clinical conditions, the 239 prevalence in LESS-CHRON is 14% vs. 8% in STOPPFrail. 240 Another relevant finding is that although the number of criteria analyzed for each tool 241 was similar (27 in LESS-CHRON and 25 in STOPPFrail), it is true that in STOPPFrail 242 only two criteria made up 65% of the total (criteria B1 and E1) and one of them affected 243 70% of the patients (criterion B1). However, in LESS-CHRON, the application of the 244 different items was more homogeneous, distributed over a greater number of different 245 criteria, and the sum of the three most applied criteria was less than 50% of the total. 246 It should be note that while this study was conducted, a new version of the STOPPFrail 247 tool was published24, which incorporates new criteria and eliminates others that had 248 become obsolete. Nevertheless, none of the previously mentioned criteria that carried 249 considerable weight in LESS-CHRON have been added in the new version. 250 Finally, it is important to mention that both tools consider criteria did not detect any 251 candidate drug to be withdrawn, with a higher number in STOPPFrail compared to LESS- 252 CHRON (15 vs 8 criteria). However, after the update to the English version of 253 STOPPFrail17, 6 of those 15 criteria have been eliminated, while 2 of them have been 254 modified. With respect to LESS-CHRON, a study is currently underway to validate the 255 criteria in a cohort of institutionalized patients and outpatients. With this study, data on 256 usability and clinical utility will be obtained that will allow the criteria to be adapted to 257 actual use needs. 258 In assessing the correlation between the clinical parameters, it is worth noting that neither 259 the number of PIM nor the STOPPFrail criteria showed any sort of correlation with frailty, 260 while there was a weak and significant correlation with the LESS-CHRON criteria 261 (28.6%, p=0.014). This may be due to the fact that the degree of frailty of the patient in 262 our cohort did not exceed the threshold for which the STOPPFrail tool was designed. 263 At the same time, prognostic assessment19 and Barthel index18 again showed a weak but 264 significant correlation with the LESS-CHRON criteria, but no correlation with the 265 STOPPFrail tool. This could be explained by the fact that the LESS-CHRON criteria 266 incorporate as part of the clinical evaluation of patients. 267 Consequently, one can deduce that the LESS-CHRON tool better adapts to the patient’s 268 clinical and prognostic context, including a greater number of criteria in older chronic 269 patients with multimorbidity. However, given the low correlation rates, it would be 270 desirable to design new studies with a larger sample focused on this aspect. 271 Limitations 272 Among the limitations of the study, we highlight that the most recent version of 273 STOPPFrail was not used. In addition, section A was removed from it in order to make a 274 more realistic comparison between deprescribing criteria. Expanding this study to a larger 275 number of patients would provide more reliable data on the correlation between the tools 276 and the clinical parameters analyzed. 277 Conclusions and Implications 278 In summary, both tools have demonstrated a great capacity to identify PIM that can be 279 deprescribed in an older polymedicated population with multimorbidity. In addition, they 280 complement each other, covering a greater number of areas of pharmacotherapy between 281 the two. Though, because STOPPFrail was designed and validated in a population with a 282 poor prognosis and great frailty, LESS-CHRON appears to have greater detection 283 potential in the subgroup of patients analyzed. It is necessary to confirm these results in 284 the context of applying a real-life criteria application. 285 Conflicts of Interest 286 The authors declare that there is no conflict of interest regarding the publication of this 287 article. 288 References 289 1. Zemedikun DT, Gray LJ, Khunti K, Davies MJ, Dhalwani NN. Patterns of 290 Multimorbidity in Middle-Aged and Older Adults: An Analysis of the UK Biobank Data. 291 Mayo Clin Proc. 2018;93(7):857-866. doi: 10.1016/j.mayocp.2018.02.012. 292 2. Chudasama YV, Khunti K, Gillies CL, Dhalwani NN, Davies MJ, Yates T, et al. 293 Healthy lifestyle and life expectancy in people with multimorbidity in the UK Biobank: 294 A longitudinal cohort study. PLoS Med. 2020;17(9):e1003332. https:// 295 doi.org/10.1371/journal.pmed.1003332. 296 3. Zazzara MB, Vetrano DL, Carfì A, Onder G. Frailty and chronic disease. Panminerva 297 Med. 2019;61(4):486-492. doi: 10.23736/S0031-0808.19.03731-5. 298 4. Hanlon P, Nicholl BI, Jani BD, Lee D, McQueenie R, Mair FS. Frailty and pre-frailty 299 in middle-aged and older adults and its association with multimorbidity and mortality: a 300 prospective analysis of 493 737 UK Biobank participants. Lancet. 2018;3(7):323–332. 301 doi: 10.1016/S2468-2667(18)30091-4. 302 5. Barnett K, Mercer SW, Norbury M, Watt G, Wyke S, Guthrie B. Epidemiology of 303 multimorbidity and implications for health care, research, and medical education: a cross- 304 sectional study. Lancet. 2012;380: 37–43. doi: 10.1016/S0140-6736(12)60240-2. 305 6. Bernabeu-Wittel M, Ollero-Baturone M, Nieto-Martín D, García-Morillo S, 306 Goicoechea-Salazar J. Patient-Centered Care for Older Adults with Multiple Chronic 307 Conditions: These are the Polypathological Patients! J Am Geriatr Soc. 2013;61(3):475- 308 6. doi: 10.1111/jgs.12142. 309 7. Bernabeu Wittel M, Barón Franco B, Murcia Zaragoza J, Fuertes Martín A, Ramos 310 Cantos C, Fernández Moyano A, et al. A multi-institutional, hospital-based assessment 311 of clinical, functional, sociofamilial and health-care characteristics of polypathological 312 patients (PP). Arch Gerontol Geriatr. 2011;53(3):284–91. doi: 313 10.1016/j.archger.2010.12.006. 314 8. Page AT, Falster MO, Litchfield M, Pearson SA, Etherton-Beer CH. Polypharmacy 315 among older Australians, 2006-2017: a population-based study. Med J Aust. 316 2019;211(2):71-75. doi: 10.5694/mja2.50244 317 9. Martínez-Sotelo J, Pinteño-Blanco M, García-Ramos R, Llobera-Cànaves J, Cadavid- 318 Torres MI. Pharmacist-led intervention on potentially inappropriate prescription in 319 patients with polypharmacy: PHARM-PC clinical trial protocol. Farm Hosp. 320 2021;45(4):210-215. English. doi: 10.7399/fh.11575. 321 10. Gallagher J, O’Sullivan D, McCarthy S, Gillespie P, Woods N, O’Mahony D, et al. 322 Structured pharmacist review of medication in older hospitalised patients: a cost- 323 effectiveness analysis. Drugs Aging. 2016; 33(4): 285–294. 324 11. Ulley J, Harrop D, Ali A, Alton S, Fowler Davis S. Deprescribing interventions and 325 their impact on medication adherence in community-dwelling older adults with 326 polypharmacy: a systematic review. BMC Geriatr. 2019;18;19(1):15. doi: 327 10.1186/s12877-019-1031-4. 328 12. Rodríguez Pérez A, Alfaro Lara ER, Nieto Martín MD, Ruiz Cantero A, Santos 329 Ramos B. Deprescribing in patients with multimorbidity: A necessary process. European 330 Journal of Internal Medicine. 2015;26(7):e18-e19. doi: 10.1016/j.ejim.2015.06.011. 331 13. Thompson W, Lundby C, Graabæk T, Nielsen DS, Ryg J, Søndergaard J et al. Tools 332 for Deprescribing in Frail Older Persons and Those with Limited Life Expectancy: A 333 Systematic Review. J Am Geriatr Soc. 2019;67(1):172-180. doi: 10.1111/jgs.15616. 334 14. Rodríguez Pérez A, Alfaro Lara ER, Albiñana Pérez S, Nieto Martín MD, Díez 335 Manglano J, Pérez Guerrero C et al. Novel tool for deprescribing in chronic patients with 336 multimorbidity: List of Evidence-Based Deprescribing for Chronic Patients criteria. 337 Geriatr Gerontol Int. 2017;17(11):2200-2207. doi: 10.1111/ggi.13062. 338 Formatted: Spanish (Spain) 15. Lavan A, Gallagher P, Parsons C et al. STOPPFrail (Screening Tool of Older Persons 339 Prescriptions in Frail adults with limited life expectancy): Consensus validation. Age 340 Ageing 2016;45:1–12. DOI: 10.1093/ageing/afx005 341 16. Curtin D, Dukelow T, James K, O'Donnell D, O'Mahony D, Gallagher P. 342 Deprescribing in multi-morbid older people with polypharmacy: agreement between 343 STOPPFrail explicit criteria and gold standard deprescribing using 100 standardized 344 clinical cases. Eur J Clin Pharmacol. 2019 Mar;75(3):427-432. doi: 10.1007/s00228-018- 345 2598-y. 346 17. Pfeiffer E. A short portable mental status questionaire for the assesment of organic 347 brain deficit in elderly patients. J Am Geriatr Soc 1975; 23: 433-441. 348 18. Mahoney FI, Barthel DW. Functional evaluation: the Barthel Index. Md State Med J 349 1965;4:61–5. 350 19. Bernabeu Wittel M, Moreno Gaviño L, Ollero Baturone M, Barón Franco B, Díez 351 Manglano J, Rivas-Cobas C, et al. Validation of PROFUND prognostic index over a 352 fouryear follow-up period. Eur J Intern Med. 2016;36:20–4. 353 20. Amblàs Novellasa J, Martori JC, Molist Brunet N, Oller R, Gómez-Batiste X, 354 Espaulella Panicot J. [Frail-VIG index: Design and evaluation of a new frailty index based 355 on the Comprehensive Geriatric Assessment].Rev Esp Geriatr Gerontol. 2017;52(3):119- 356 127. doi: 10.1016/j.regg.2016.09.003. 357 21. Hulley SB, Cummings SR, Browner WS, Grady D, Newman TB. Designing clinical 358 research: an epidemiologic approach. 4th ed. Philadelphia, PA: Lippincott Williams & 359 Wilkins. 2013;6C:79. 360 Supplementary Material 2 Click here to access/download Supplementary Material Supplementary material 2.docx Supplementary Material 3 Click here to access/download Supplementary Material Supplementary material 3..docx Supplementary Material 4 Click here to access/download Supplementary Material Supplementary material 4.docx