Journals of Gerontology: Medical Sciences cite as: J Gerontol A Biol Sci Med Sci, 2022, Vol. 77, No. 1, 148–154 doi:10.1093/gerona/glab102 Advance Access publication April 22, 2021 Research Article Mortality, Falls, and Fracture Risk Are Positively Associated With Frailty: A SIDIAP Cohort Study of 890 000 Patients Robert Middleton, BMBCh, MA(Cantab), MRCS,1 Jose Luis Poveda, Msc,2 Francesc Orfila Pernas, MD, PhD,3,4 Daniel Martinez Laguna, MD, PhD,2,4 Adolfo Diez Perez, MD, PhD,2,5, Xavier Nogués,MD, PhD,2,6 Cristina Carbonell Abella, MD, PhD,3,7 Carlen Reyes, MD, PhD,2,4 and Daniel Prieto-Alhambra, MD, MSc, PhD1,2,* Nuffield Department of Orthopaedics, Rheumatology and Musculoskeletal Sciences, Botnar Research Centre, University of Oxford, 1 UK. 2CIBERFes, Instituto Carlos III, Madrid, Spain. 3Gerència Territorial de Barcelona, Institut Català de la Salut, Spain. 4Fundació Institut Universitari per a la recerca a l’Atenció Primària de Salut Jordi Gol i Gurina (IDIAPJGol), Barcelona, Spain. 5Musculoskeletal Research Unit, IMIM-Hospital del Mar, Barcelona, Spain. 6Internal Medicine Department, IMIM-Hospital del Mar, Barcelona, Spain. 7Universitat Autònoma de Barcelona, Bellaterra (Cerdanyola del Vallès), Spain *Address correspondence to: Daniel Prieto-Alhambra, MD, MSc, PhD, Nuffield Department of Orthopaedics, Rheumatology and Musculoskeletal Sciences, Botnar Research Centre, University of Oxford, Windmill Road, OX3 7LD Oxford, UK. E-mail: Daniel.prietoalhambra@ndorms.ox.ac.uk Received: August 22, 2020; Editorial Decision Date: March 28, 2021 Decision Editor: Anne B. Newman, MD, MPH, FGSA Abstract Background: Frail subjects are at increased risk of adverse outcomes. We aimed to assess their risk of falls, all-cause mortality, and fractures. Method: We used a retrospective cohort study using the Sistema d’Informació per al Desenvolupament de l’Investigació en Atenció Primària database (>6 million residents). Subjects aged 75 years and older with ≥1 year of valid data (2007–2015) were included. Follow-up was carried out from (the latest of) the date of cohort entry up to migration, end of the study period or outcome (whichever came first). The eFRAGICAP classified subjects as fit, mild, moderate, or severely frail. Outcomes (10th revision of the International Classification of Diseases) were incident falls, fractures (overall/hip/vertebral), and all-cause mortality during the study period. Statistics: hazard ratios (HRs), 95% CI adjusted (per age, sex, and socioeconomic status), and unadjusted cause-specific Cox models, accounting for competing risk of death (fit group as the reference). Results: A total of 893 211 subjects were analyzed; 54.4% were classified as fit, 34.0% as mild, 9.9% as moderate, and 1.6% as severely frail. Compared with the fit, frail had an increased risk of falls (adjusted HR [95% CI] of 1.55 [1.52–1.58], 2.74 [2.66–2.84], and 5.94 [5.52–6.40]), all-cause mortality (adjusted HR [95% CI] of 1.36 [1.35–1.37], 2.19 [2.16–2.23], and 4.29 [4.13–4.45]), and fractures (adjusted HR [95% CI] of 1.21 [1.20–1.23], 1.51 [1.47–1.55], and 2.36 [2.20–2.53]) for mild, moderate, and severe frailty, respectively. Severely frail had a high risk of vertebral (HR of 2.49 [1.99–3.11]) and hip fracture (HR [95% CI] of 1.85 [1.50–2.28]). Accounting for competing risk of death did not change results. Conclusion: Frail subjects are at increased risk of death, fractures, and falls. The eFRAGICAP tool can easily assess frailty in electronic primary care databases in Spain. Keywords: Epidemiology, Frailty, Risk factors The global population is increasing in size and age; United Nations ceived increasing attention in recent years; the early identification of estimates that 1 in 6 individuals will be older than 65 years by 2050, frail subjects would help design specific preventive strategies for this rising to 1 in 4 for North America and Europe (1). Aging popula- population at risk. tions are expected to place increased pressure on health care systems, Frailty is a concept that recognizes deterioration of multiple which will need to support a greater number of patients with chronic body systems and failure of homeostatic mechanisms. Frail individ- conditions. Frailty, as a measure of vulnerability to ill health, has re- uals are at greater risk of adverse outcomes (higher mortality, loss © The Author(s) 2021. Published by Oxford University Press on behalf of The Gerontological Society of America. 148 This is an Open Access article distributed under the terms of the Creative Commons Attribution License (https:// creativecommons.org/licenses/by/4.0/), which permits unrestricted reuse, distribution, and reproduction in any medium, provided the original work is properly cited. Journals of Gerontology: MEDICAL SCIENCES, 2022, Vol. 77, No. 1 149 of independence, complications, and prolonged recovery), and these origin, socioeconomic status), clinical measurements (height, weight, may result from relatively minor insults (2–4). Identification of frailty waist circumference, blood pressure), recorded diagnoses gathered can be achieved with phenotypic models or cumulative deficit models through 10th revision of the International Classification of Diseases (2,5). Phenotypic frailty is determined by clinical assessment for (ICD-10) codes as well as pharmacy invoicing data and hospital ad- the presence of signs such as slow gait, weak grip, or weight loss. missions (17). Cumulative deficit models determine frailty by the presence of defined conditions, as identified in the models’ development. These cumulative Study Period, Participants, Inclusion/Exclusion deficit models generate a frailty index, with higher values indicating Criteria, and Follow-up greater degrees of frailty. Compared to the phenotypic assessment of The study period was from January 1, 2007 to the December 31, frailty, frailty indexes can be implemented in electronic health care 2015. Eligible patients were all subjects of at least 75 years old records (EHRs), are less time-consuming, and could be used to screen during the study period with at least 1 year of valid data previous to for frailty at a population level (6). The electronic Frailty Index (eFI) the eFRAGICAP calculation (index date). There were no exclusion was developed using primary care data in the United Kingdom and criteria. Follow-up was carried out from the latest of date of cohort incorporates 36 potential deficits. Increasing levels of frailty (categor- entry (whenever the subjects were eligible) or the start of the study ized as mild, moderate, and severe) demonstrated increased mortality, period, until transfer out of the catchment area, end of the study hospital admission, and nursing home admission risks (7). The eFI period or study outcome (death, fractures) whichever came first. has subsequently been modified to create the eFRAGICAP for com- patibility with Catalonian primary care data, with validation work Exposure confirming good discriminatory capacity (8). Subjects’ frailty was the main exposure which was calculated for Frailty prevalence is dependent on the population being studied all eligible subjects at the index date with the eFRAGICAP tool and the frailty assessment method used. Collard et al. reported a (8). The eFRAGICAP is a validated and adapted tool based on the wide range (4.0%–59.1%) in their 2012 systematic review, which in- eFI from the United Kingdom (7), which comprises a total of 36 cluded multiple frailty measures across countries (with cohorts ran- deficits based on the Rockwood model of frailty including medical ging in size from 230 to 8 914) (9). In the development of the eFI, an and pharmacy information (ICD-10, ATC codes, etc.) registered in external validation using 516 007 subjects from the United Kingdom primary care records. Categories of fit, mild frailty, moderate frailty, demonstrated a frailty prevalence of 37%, 16%, and 4% for mild, and severe frailty were defined by quartiles using the 99th centile as moderate, and severe frailty, respectively (7). Despite variations in the upper limit as in the eFI study (7). All individual deficit, deficit frailty prevalence, a meta-analysis of 19 studies (ranging in size from count, and frailty categories were determined for the study popula- 754 to 36 306, across multiple countries) reported increased mor- tion (Supplementary Table 1). tality risk with increasing frailty index (10). Fragility fractures are fractures occurring following low-energy trauma and are associated with increased morbidity and mortality Study Outcomes (11–13). Management of such injuries is a growing challenge for The outcomes of interest were incident falls, all-cause mortality, and health care systems, with significant economic consequences. For incident fractures: (i) all fractures (excluding face, skull, and digits), 2005, the cost of managing osteoporosis-related fractures in the (ii) vertebral fractures, and (iii) hip fractures. All outcomes were as- United States was estimated at $17 billion, and for the United certained during the study period and identified using ICD-10 codes Kingdom in 2000 this figure was $1.8 billion, and these figures are (Supplementary Table 2). expected to rise (14,15). In a study of 6 724 females aged 65 years and older, increasing phenotypic frailty was associated with in- Confounders creased risk of hip fracture, non-spine fracture, and mortality with a Confounders associated with both the risk of fractures and frailty 9-year follow-up (16). There is a scarcity of reports using large na- were predefined a priori. These were the age of the subject at the tional data sets investigating associations between fragility fractures index date, sex, and socioeconomic status measured with the and frailty. Given the increasing use of frailty measures to identify MEDEA index tool (18). vulnerable patients, and the growing population at risk of fragility fractures, a contemporary investigation of these factors is of value. Ethical Considerations The aim of this study was to describe the prevalence of frailty This study was approved on the February 28, 2018 by the Scientific among old subjects in Catalonia using the eFRAGICAP, and to deter- and Ethical Committees of SIDIAP (DE-020-Certificat CEIC project mine associations between mortality, fracture risk, and frailty using a nº P17/209). large national health care database. Statistical Analyses Baseline characteristics of participants were described using mean Materials and Methods and standard deviation, or median and interquartile range if the Study Design, Data Source, and Population distribution of the data was asymmetric, for continuous variables We conducted a retrospective cohort study using data from the and frequencies for categorical variables. Potential confounders Sistema d’Informació per al Desenvolupament de l’Investigació en considered for adjustment were those known relevant for frailty Atenció Primària (SIDIAP) database (http://www.sidiap.org). The and fractures such as age, sex, and socioeconomic status. These SIDIAP database comprises anonymized electronic medical rec- confounders were measured at baseline. The Cox proportional ords of over 5 million patients attending primary care centers in hazards regression models were fitted to estimate hazard ratios Catalonia (>80% of the population). Information collected included (HRs) and 95% confidence intervals (95% CIs) adjusted for age, sociodemographic information (year of birth, gender, country of sex, and socioeconomic status taking the fit group as the reference 150 Journals of Gerontology: MEDICAL SCIENCES, 2022, Vol. 77, No. 1 group. Adjusted and unadjusted cause-specific HRs were calcu- Cox regression models for the unadjusted and adjusted asso- lated using cause-specific Cox models for the risk of fractures (all, ciation between frailty and the outcomes of interest (death, frac- hip, and vertebra) and the risk of falls accounting for each event tures, and falls) are reported in Table 2. After adjusting for potential for the competing risk of death. Statistical significance was defined confounders and compared with the fit subjects (reference group), at the p <.05 level. All statistical analyses were conducted using the risk of falling nearly doubled with every increase in frailty (ad- R version 3.4.5 for Windows using the mstate package and SPSS justed HR of 1.55, 2.74, and 5.94 for mild, moderate, and severe version 22. frailty, respectively). Mortality also showed the same increasing trend, with moderate and severe frailty reporting 2- and 4-fold in- creased risk of death compared to the fit group. Regarding fractures, Results overall fracture risk also increased with increasing frailty (adjusted Baseline characteristics of the population analyzed are reported in HR of 1.21, 1.51, and 2.36 for mild, moderate, and severe frailty, Table 1. A total of 893 211 subjects were included and analyzed. respectively). This was also true for individual fracture categories; Most were old women (mean age of 79 years), rather healthy (with 0 compared to fit subjects, subjects with severe frailty seemed to have or 1 deficits) coming from an urban setting. The population studied a greater risk of vertebral and hip fracture (adjusted HR of 2.49 and was evenly distributed across the socioeconomic strata. When cal- 1.85, respectively). Cause-specific Cox analysis accounting for the culating the frailty index with the eFRAGICAP tool, the majority of competing risk of death (Supplementary Tables 3 and 4) were in line subjects included were classified as fit or having mild frailty (54% with the previous results. and 34%, respectively). Sensitivity analyses were carried out in men and women separ- Patients were followed for 7.23 years (95% CI: 7.23–7.24), 8.30 ately; adjusted Cox regression models found an overall higher risk (8.30–8.31), and 8.63 (8.63-8.63) for the outcomes of mortality, of suffering all outcomes (death, fractures, and falls) with increasing fractures, and falls. Incidence rates per 1 000 person-year (95% CI) frailty compared to the fit category in both men and women. Men was 55.82 (55.33–56.07) for mortality, 18.81 (18.53–18.96) for with frailty (mild, moderate, and severe) had a higher risk of death, fractures, and 10.98 (10.76–11.09) for falls. fractures, and falls compared to women (Supplementary Table 5). Kaplan–Meier survival curves for the outcomes of mor- Cause-specific Cox analysis accounting for the competing risk of tality, fractures, and falls are reported in Figures 1–3. During the death (Supplementary Table 6) did not change the results. follow-up, mortality showed the greatest increasing trend with increasing frailty (especially pronounced for those classified as moderate or severe frailty) compared to those with mild or those Discussion classified as fit (Figure 1). To a lesser degree, the same trend is seen In this large longitudinal population-based cohort study, we found for overall fractures (Figure 2; Supplementary Figure 1) and falls that increasing levels of frailty (defined by eFRAGICAP) were asso- (Figure 3) with a greater incidence of outcomes among those with ciated with increased risks of mortality, fractures, and falls. Subjects moderate and severe frailty. with moderate and severe frailty had over 2- and 4-fold increased Table 1. Baseline Characteristics of the Population Analyzed Characteristics Sex Total Female Male N (%) 893 211 (100) 533 457 (59.73) 359 754 (40.27) Age (y) Mean (SD) 78.90 (5.23) 79.38 (5.54) 78.17 (4.65) Median (IQR) 76.08 (75−81.75) 76.75 (75–82.67) 75.17 (75–80.33) No. of accumulated deficits Median (IQR) 1 (0–3) 1 (0–3) 1 (0–3) Frailty categorya, n (%) Fit 486 683 (54.49) 291 415 (54.63) 195 268 (54.28) Mild 303 802 (34.01) 179 578 (33.66) 124 224 (34.53) Moderate 88 473 (9.91) 53 491 (10.03) 34 982 (9.72) Severe 14 253 (1.60) 8 973 (1.68) 5 280 (1.47) Habitat, n (%) Rural 181 731 (20.35) 104 933 (19.67) 76 798 (21.35) Urban 709 660 (79.45) 427 337 (80.11) 282 323 (78.48) Missing 1 820 (0.20) 1 187 (0.22) 359 121 (0.18) MEDEAb, n (%) U1 138 565 (19.53) 86 554 (20.25) 52 011 (18.42) U2 117 648 (16.58) 71 109 (16.64) 46 539 (16.48) U3 112 573 (15.86) 66 899 (15.65) 45 674 (16.18) U4 106 135 (14.96) 62 153 (14.54) 43 982 (15.58) U5 91 270 (12.86) 53 395 (12.49) 37 875 (13.42) Missing 143 469 (20.22) 87 227 (20.41) 56 242 (19.92) Notes: IQR = interquartile range. a Fit = 0 or 1 deficit, mild = 2–4 deficits, moderate = 5–8 deficits, and severe = ≥9 deficits. bMEDEA = social deprivation score being U1 the least deprived and U5 the most deprived. Journals of Gerontology: MEDICAL SCIENCES, 2022, Vol. 77, No. 1 151 risk of death compared with those classified as fit. Regarding the risk of fractures, and compared to fit subjects, those with moderate and severe frailty had an increased risk of vertebral fractures (HR of 1.67 and 2.49, respectively) and hip fractures (HR 1.32 and 1.85). Total eFRAGICAP deficit count also demonstrated a positive associ- ation with mortality and fracture risk. Many frailty instruments have been developed in recent years; however, few are the ones that assessed frailty on EHRs (7,19–22). The prevalence of frailty within the Catalonian population, as de- fined by the eFRAGICAP, was similar to that of the UK population, as defined by the eFI in Clegg et al.’s validation study. Of the study populations, the majority were predominantly fit or mildly frail (80% for eFI and 89% for eFRAGICAP), with moderate (16% vs 10%, respectively) and severe frailty (4% vs 2%, respectively) less prevalent (7). Recently, the eFI was also adapted to an EHR in the United States and calculated for 12 798 patients (22), with 40.1% of their patients’ classified as frail, which is slightly lower than our re- sults. The adaptation of the eFI in each of the studies, with different Figure 1. Nine-year Kaplan–Meier survival curve for the outcome of mortality codes included depending on their availability in the databases, according to the variable frailty. could explain the different prevalence of frailty found. Further dif- ferences with previous reports such as the customized frailty indexes used or the lower number of patients included (19,21) render com- parison difficult. Previous studies carried out in Catalonia (23,24), both using Fried’s criteria in subjects older than 70 and 75 years, demonstrated different prevalence’s for frailty, with 31% frail and 49% prefrail in the Serra-Prats et al.’s investigation (23) and 46.9% prefrailty and 9.6% of frailty in the Escobar-Bravo et al.’s investigation (24). While phenotypic and index-based frailty measures cannot be directly com- pared, these results are not dissimilar. Regarding mortality, our estimations were similar to that re- ported by Clegg et al. (7). At 5 years, the adjusted mortality HR was reported as 1.66, 2.54, and 3.84 for mild, moderate, and severe frailty, respectively. Our population, with 9-year follow-up, demon- strated similar HRs of 1.36, 2.19, and 4.29 respectively. Two of the reports that assessed frailty in EHRs (19,22) and a previous meta- analysis (10) found that frailty was associated with an increased risk of death and another one with an overall risk of adverse outcomes (21), which is also in accordance with our results. Figure 2. Nine-year Kaplan–Meier survival curve for the outcome of fracture Concerning fracture risk, Ensrud et al. (16) investigated the asso- according to the variable frailty. ciation between phenotypic frailty and fractures in women aged at least 65 years during 9 years of follow-up and found similar results as ours, despite the different frailty definition used (phenotypic vs cumulative health deficits). We found an adjusted HR of 1.32 for hip fracture in the moderately frail cohort, and an adjusted HR of 1.27 with intermediate frailty. Risk was greater in the severely frail group, at 1.85 versus 1.4 in Ensrud’s cohort (16). Methods of frailty determination, and variables adjusted for, likely contributed to the differences here. Kojima’s meta-analyses of frailty and fracture (any site) risks are also in keeping with these figures (HR of 1.57 for frail and 1.30 for prefrail) (25). Other previous studies that analyzed the association between frailty and vertebral fractures found similar re- sults to ours; a 30% increased risk of vertebral fractures with every 0.10 increase in the frailty index score was reported in the Canadian Multicentre Osteoporosis Study (CaMos study) and this risk was higher for vertebral than for hip fractures (26). As the aforemen- tioned study, we also found an increased risk of vertebral and hip fractures in our severely frail population. However, the increased risk found for vertebral fractures needs to be interpreted with cau- tion, given that we only gathered clinical vertebral fractures (not al- Figure 3. Nine-year Kaplan–Meier survival curve for the outcome of falls ways radiologically confirmed), unlike hip fractures which usually according to the variable frailty. require hospital care. Vertebral fractures are often underdiagnosed 152 Journals of Gerontology: MEDICAL SCIENCES, 2022, Vol. 77, No. 1 (27) and therefore medical doctors may be more prone to search (reference) (reference) for them among frail subjects rather than in those who are fit. At 1.541.60 2.702.89 5.706.59 1.211.22 1.521.58 2.662.84 5.526.40 1.211.22 95% CI last, increasing frailty implies a greater number of comorbidities, which could increase the risk of vertebral fractures on their own and require closer monitoring by the medical doctor with a greater number of visits, leading to a greater number of diagnoses of verte- 1.00 1.57 2.80 6.13 1.22 1.00 1.55 2.74 5.94 1.21 HR bral fractures. To our knowledge, this is the first study to use an eFI such as the Number of eFRAGICAP in a large real-world database in Spain. The previous 28 593 18 508 4 332 757 — — — — — Events longitudinal studies (23,24) carried out in Spain used the Frieds’ cri- Falls teria which require large time-consuming questionnaires which are difficult to implement in primary care. Moreover, our study popula- (reference) (reference) tion is one of the largest to date used for the investigation of a frailty 0.971.05 1.051.23 1.201.82 1.011.03 1.061.15 1.221.43 1.502.28 1.041.07 95% CI index and associations with mortality and fracture risk. The SIDIAP database is a primary care database designed specifically for use as a research database with data quality validation measures, and we 1.00 1.00 1.14 1.48 1.02 1.00 1.11 1.32 1.85 1.05 HR have used a frailty index that has been previously validated for use in Hip Fracture the population of interest which supports the external validity of our Number of results and the use of this measure in real world. This is reinforced Events 7 504 3 559 680 89 — — — — — given that our results are similar to previously published figures for the outcomes of interest despite the population differences. We have demonstrated an increased risk of mortality, fracture, (reference) (reference) 1.221.37 1.662.00 2.283.56 1.101.13 1.171.29 1.521.83 1.993.11 1.091.11 and falls in association with increasing degrees of frailty, as deter- 95% CI mined using the eFRAGICAP tool. These risks remained after ad- justment for confounding variables. Frailty assessment is a valuable method to identify at-risk patients; however, such assessments need Vertebral Fracture 1.00 1.29 1.82 2.85 1.12 1.00 1.23 1.67 2.49 1.10 HR to be easy to implement to be able to use them in primary care where time per visit is limited. A frailty tool such as eFRAGICAP which Number of can categorize frailty using routinely collected primary care data (in Events 3 812 2 293 535 81 — — — — — comparison to phenotypic assessments requiring specific clinical re- views) is of great relevance. The same is seen in the United Kingdom, where frailty measures (eFI) are being actively implemented in pri- (reference) (reference) mary care in the United Kingdom in all patients over the age of 65. 1.121.15 1.311.38 1.842.12 1.061.07 1.201.23 1.471.55 2.202.53 1.081.09 95% CI This facilitates targeted interventions to maintain patient health, re- ducing the risk of unplanned hospital admissions and death. The associations between frailty and unplanned hospital admission and Table 2. Association Between Frailty and the Risk of Death, Fractures, and Falls mortality supported the implantation of this process. This study also 1.00 1.14 1.35 1.97 1.06 1.00 1.21 1.51 2.36 1.09 HR adds to the value of such assessments, as frailty level can also be Fractures (all) used to identify patients at increased risk of fracture (with the at- Number of tendant morbidity and mortality). Fracture prediction tools (such 486 683 303 802 88 473 14 253 — — — — — Events as the Fracture Risk Assessment Tool—FRAX) are already used to identify patients who may benefit from bone protection therapy. Li et al. compared a frailty index and the FRAX in predicting fu- (reference) (reference) ture risk of fractures, and found similar performance, particularly 1.181.20 1.691.75 2.682.88 1.091.10 1.351.37 2.162.23 4.134.45 1.151.16 95% CI for high-risk patients (28). The identification of frail patients can, therefore, also be seen as an opportunity to start bone protection, to reduce fracture risk. 1.00 1.19 1.72 2.77 1.10 1.00 1.36 2.19 4.29 1.16 From our results, we can speculate that any strategy for redu- HR cing the impact of fractures should include not only bone-targeted Number of interventions but also general management of the frailty syndrome. Mortality 162 112 87 412 19 620 2 972 — — — — — Moreover, since fractures can be the consequence but also the cause Events of increased frailty, we cannot disentangle both problems in the clin- ical and public health approach to these patients. Note: HR = hazard ratio. Limitations of our study warrant consideration. In the first place, the identification of frail subjects is complex and involves medical, social, Total deficit count Total deficit count Fit (ref. p < .001) Fit (ref. p < .001) and phenotypical aspects that are not always captured in an electronic Moderate frailty Moderate frailty database such as the SIDIAP database. Despite that our results with the Severe frailty Severe frailty Mild frailty Mild frailty Nonadjusted eFRAGICAP are similar to what has been previously reported in the United Kingdom (7), we cannot exclude the possibility that we might Adjusted have underestimated the true prevalence of subjects in each frailty cat- egory. Second, we did not fully adjust for other potential confounders, Journals of Gerontology: MEDICAL SCIENCES, 2022, Vol. 77, No. 1 153 such as smoking, alcohol consumption, or physical activity. These were 3. De Lepeleire J, Iliffe S, Mann E, Degryse JM. Frailty: an emerging con- indirectly captured through the deficits of the eFRAGICAP and there- cept for general practice. Br J Gen Pract. 2009;59(562):e177–e182. fore partially accounted for. Moreover, frail subjects require usually a doi:10.3399/bjgp09X420653 4. Sieber CC. Frailty—from concept to clinical practice. Exp Gerontol. closer monitorization because of their comorbidities and therefore are 2017;87(Pt B):160–167. doi:10.1016/j.exger.2016.05.004 more prone to visit the primary care doctors, increasing the possibil- 5. Cesari M, Gambassi G, van Kan GA, Vellas B. The frailty phenotype and ities of being diagnosed of an outcome; this could partly explain our the frailty index: different instruments for different purposes. Age Ageing. increased risk of vertebral fractures. Furthermore, we could only as- 2014;43:10–12. doi:10.1093/ageing/aft160 sess clinical vertebral fractures (with no information regarding x-rays), 6. Kim DH. Measuring frailty in health care databases for clinical care and re- leading to a possible underestimation of the true incidence of vertebral search. Ann Geriatr Med Res. 2020;24:62–74. doi:10.4235/agmr.20.0002 fractures. At last, in spite that the collection of data in the SIDIAP data- 7. Clegg A, Bates C, Young J, et al. Development and validation of an elec- base has been previously validated (17), there is still the possibility of tronic frailty index using routine primary care electronic health record misclassifications, which should be considered. data [published correction appears in Age Ageing. 2017 Jan 17]. Age Ageing. 2016;45:353–360. doi:10.1093/ageing/afw039 8. Reyes C, Orfila Pernas F, Prieto-Alhambra D. Validation of an electronic index of frailty in primary care (EFRAGICAP). World Congress on Conclusion Osteoporosis, Osteoarthritis and Musculoskeletal Diseases (WCO-IOF- ESCEO 2020): Poster Abstracts. Osteoporosis Int 2020;31:133–621. Frailty measured by the eFRAGICAP tool is associated with an increased doi:10.1007/s00198-020-05696-3 risk of mortality, fractures, and falls in subjects over the age of 75 years. 9. Collard RM, Boter H, Schoevers RA, Oude Voshaar RC. Prevalence of The eFRAGICAP tool can be easily implemented for the assessment of frailty in community-dwelling older persons: a systematic review. J Am frailty in routinely collected primary care databases in Spain. Geriatr Soc. 2012;60:1487–1492. doi:10.1111/j.1532-5415.2012.04054.x 10. Kojima G, Iliffe S, Walters K. Frailty index as a predictor of mortality: a systematic review and meta-analysis. Age Ageing. 2018;47:193–200. doi:10.1093/ageing/afx162 Supplementary Material 11. Bliuc D, Nguyen ND, Milch VE, Nguyen TV, Eisman JA, Center JR. Supplementary data are available at The Journals of Gerontology, Mortality risk associated with low-trauma osteoporotic fracture and sub- Series A: Biological Sciences and Medical Sciences online. sequent fracture in men and women. J Am Med Assoc. 2009;301:513–521. doi:10.1001/jama.2009.50 12. Nazrun AS, Tzar MN, Mokhtar SA, Mohamed IN. A systematic review of the outcomes of osteoporotic fracture patients after hospital discharge: Funding morbidity, subsequent fractures, and mortality. Ther Clin Risk Manag. This work was supported by the National Institute for Health Research 2014;10:937–948. doi:10.2147/TCRM.S72456 (NIHR) Oxford Biomedical Research Centre (BRC). D.P.-A. is funded through 13. Johnell O, Kanis JA. An estimate of the worldwide prevalence and disability a NIHR Senior Research Fellowship (grant number SRF-2018-11-ST2-004). associated with osteoporotic fractures. Osteoporos Int. 2006;17:1726– The study was also supported by the Centro de Investigación Biomédica en 1733. doi:10.1007/s00198-006-0172-4 Red sobre Fragilidad y Envejecimiento Saludable (CIBERFES). The views ex- 14. Burge R, Dawson-Hughes B, Solomon DH, Wong JB, King A, Tosteson A. pressed in this publication are those of the author(s) and not necessarily those Incidence and economic burden of osteoporosis-related fractures in of the NHS, the NIHR, or the Department of Health. the United States, 2005–2025. J Bone Miner Res. 2007;22:465–475. doi:10.1359/jbmr.061113 15. Burge R, Worley D, Johansen A, Bhattacharyya S, Bose U. The cost of Conflict of Interest osteoporotic fractures in the UK: projections for 2000–2020. J Med Econ. 2001;4:1–4, 51–62. doi:10.3111/200104051062 A.D.P. reports personal fees from Amgen, UCB, Gilead, and Theramex, and 16. Ensrud KE, Ewing SK, Taylor BC, et al. Frailty and risk of falls, fracture, and other from Active Life Sci, outside the submitted work; D.M.L. reports per- mortality in older women: the Study of Osteoporotic Fractures. J Gerontol sonal fees from AMGEN, ITALFARMACO, RUBIO, and NOVARTIS, outside A Biol Sci Med Sci. 2007;62:744–751. doi:10.1093/gerona/62.7.744 the submitted work; D.P.-A. reports grants and other from AMGEN, grants, 17. Garcia-Gil MM, Hermosilla E, Prieto-Alhambra D, et al. Construction and nonfinancial support, and other from UCB Biopharma, and grants from Les validation of a scoring system for the selection of high-quality data in a Laboratoires Servier, outside the submitted work; Janssen, on behalf of IMI- Spanish population primary care database (SIDIAP). Inform Prim Care funded EHDEN and EMIF consortiums, and Synapse Management Partners 2012;19:135–45. doi:10.14236/jhi.v19i3.806 have supported training programs organized by D.P.-A.’s department and open 18. Reyes C, García-Gil M, Elorza JM, et al. Socioeconomic status and for external participants. X.N. works as a consultant and has participated in its association with the risk of developing hip fractures: a region- lectures for AMGEN, UCB, and Lilly. C.C.A., C.R., R.M., and F.O.P. have wide ecological study. Bone. 2015;73:127–131. doi:10.1016/j. nothing to disclose. bone.2014.12.019 19. Lekan DA, Wallace DC, McCoy TP, Hu J, Silva SG, Whitson HE. Frailty Acknowledgment assessment in hospitalized older adults using the electronic health record. Biol Res Nurs. 2017;19:213–228. doi:10.1177/1099800416679730 The authors would like to acknowledge the SIDIAP team for their support. 20. Anzaldi LJ, Davison A, Boyd CM, Leff B, Kharrazi H. Comparing clin- Authors Carlen Reyes and Daniel Prieto-Alhambra are join senior authorship. ician descriptions of frailty and geriatric syndromes using electronic health records: a retrospective cohort study. BMC Geriatr. 2017;17:248. References doi:10.1186/s12877-017-0645-7 1. United Nations, Department of Economic and Social Affairs, Population 21. Drubbel I, de Wit NJ, Bleijenberg N, Eijkemans RJ, Schuurmans MJ, Division. World Population. Prospects 2019: Highlights (ST/ESA/ Numans ME. Prediction of adverse health outcomes in older people using SER.A/423). United Nations; 2019. a frailty index based on routine primary care data. J Gerontol A Biol Sci 2. Clegg A, Young J, Iliffe S, Rikkert MO, Rockwood K. Frailty in Med Sci. 2013;68:301–308. doi:10.1093/gerona/gls161 elderly people. Lancet. 2013;381(9868):752–762. doi:10.1016/ 22. Pajewski NM, Lenoir K, Wells BJ, Williamson JD, Callahan KE. Frailty S0140-6736(12)62167-9 screening using the electronic health record within a Medicare accountable 154 Journals of Gerontology: MEDICAL SCIENCES, 2022, Vol. 77, No. 1 care organization. J Gerontol A Biol Sci Med Sci. 2019;74:1771–1777. 26. Kennedy CC, Ioannidis G, Rockwood K, et al. A frailty index predicts doi:10.1093/gerona/glz017 10-year fracture risk in adults age 25 years and older: results from the 23. Serra-Prat M, Sist X, Saiz A, et al. Clinical and functional characterization Canadian Multicentre Osteoporosis Study (CaMos). Osteoporos Int. of pre-frailty among elderly patients consulting primary care centres. J Nutr 2014;25:2825–2832. doi:10.1007/s00198-014-2828-9 Health Aging. 2016;20:653–658. doi:10.1007/s12603-016-0684-3 27. Capdevila-Reniu A, Navarro-López M, López-Soto A. Osteoporotic ver- 24. Escobar-Bravo MÁ, Jürschik P, Botigué T, Nuin C. La fragilidad como tebral fractures: a diagnostic challenge in the 21st century. Published on- predictora de mortalidad en una cohorte de edad avanzada [Frailty as a line December 3, 2019. Rev Clin Esp. 2019;S0014-2565(19)30240-1. predictor of mortality in a cohort of people aged 75 years and older]. Gac doi:10.1016/j.rce.2019.09.006 Sanit. 2014;28:489–491. doi:10.1016/j.gaceta.2014.06.005 28. Li G, Thabane L, Papaioannou A, Adachi JD. Comparison between frailty 25. Kojima G. Frailty as a predictor of fractures among community-dwelling index of deficit accumulation and fracture risk assessment tool (FRAX) in older people: a systematic review and meta-analysis. Bone. 2016;90:116– prediction of risk of fractures. Bone. 2015;77:107–114. doi:10.1016/j. 122. doi:10.1016/j.bone.2016.06.009 bone.2015.04.028