Accidental and Intentional Drug Poisoning Mortality in Older Adults in Spain: Nationwide Trends, 2001-2022 Authors: Marta Donat, Cristina Ramírez-Alcázar, Florencia Giné, Luis Sordo, Santiago Martínez Torre, María José Belza, Gregorio Barrio, Julieta Politi Published: 2026 ABSTRACT ObjectiveTo estimate time trends in fatal drug poisoning among older adults in Spain during 2001-2022 according to intent, sex, and age group.Patients and MethodsRepeated cross-sectional study from January 1, 2001, through December 31, 2022. Fatal drug poisonings included poisoning deaths from psychoactive and therapeutic substances (excluding alcohol) and were classified using International Classification of Diseases, 10th Revision, codes as accidental (X40-X44) or intentional (X60-X64). Age-standardized mortality rates per million person-years were calculated. Time trends were characterized with annual percentage changes (APCs) using Joinpoint regression. Results were stratified by sex, age (65-74, 75-84, and ≥85 years), and intent.ResultsAmong older adults, age-standardized mortality rates from drug poisoning were similar among men (21.7) and women (19.9), increased markedly with age, reaching 63.3 after age 84 years, and were higher for accidental (15.8) than intentional poisoning (4.8). Joinpoint regression shows a 2015 trend change in accidental poisoning rate from strongly upward (APC, 19.6%; 95% CI, 17.2-23.9) to moderately downward (APC, −7.4%; 95% CI, −11.3 to −4.2), whereas intentional poisoning rate rose steadily throughout the study period (APC, 7.3%). Furthermore, it shows substantial differences in time trends between age groups, but not between sexes.ConclusionDrug poisonings among older adults in Spain were predominantly accidental and rose sharply until 2015 before declining, while intentional self-poisoning increased throughout; burden was greatest in adults aged ≥85 years. Drug poisoning mortality is a growing public health problem worldwide, including North America and Europe.1, 2, 3, 4 This increase is not limited to young or middle-aged people. Older adults face a growing epidemic of drug poisonings, with rates tripling in the United States between 2002 and 2021.5 In Europe, during 2000 to 2018, poisoning mortality rates grew faster among older adults than people aged 15 to 64 years.6 In addition, older women have different patterns of intoxication than men, and the lethality of intoxication increases among older women.6 Older people are particularly vulnerable to fatal drug poisoning because of several factors, including age-related physiological changes, comorbidity, frailty, polypharmacy (estimated at 36.2% in people over 65 years) and susceptibility to medication error.7, 8, 9, 10 This excess risk is not limited to licit or illicit psychoactive substances; nonpsychoactive medications, such as those for chronic pain or other health conditions, also contribute significantly to poisoning risk.11, 12, 13 Therefore, several countries have begun to include all prescription drugs and biological substances (such as immunological therapies or blood derivatives) in their monitoring of poisoning mortality.14, 15, 16 The 2 main categories of International Classification of Diseases, 10th Revision (ICD-10), for poisoning deaths are accidental and intentional self-inflicted.17 The first are those in which death is an unintended result of substance use, whereas the second correspond to suicides. This distinction is fundamental in public health because risk factors, sociodemographic profiles, and prevention strategies differ significantly between the 2 groups. Therefore, various authors advocate classifying acute drug poisonings by intent, combining illicit psychoactive substances with medicines and therapeutic substances.18,19 In Spain, previous studies on poisoning mortality by intent among people aged 65 years or older are scarce. It has been indicated that accidental poisonings are predominant among older people, without differentiating them by age subgroups.6 Furthermore, after the Coronavirus disease 2019 pandemic, the prescription of psychotropic drugs has substantially increased, which could have altered the evolution of these deaths.9 Recent evidence shows that poisoning mortality in Spain increased markedly among adults aged 15 to 64 years during 2001 to 2022, especially after 2018.20 However, no study has examined whether these trends extend to older adults or differ by intent, age group, or sex. The objective of this study was to describe the evolution of poisoning mortality among people aged 65 years or older (older adults) in Spain during 2001 to 2022, by intent, sex, and age group and to discuss possible explanatory factors. Patients and Methods Design, Study Population, Data Sources, and Variables The evolution of drug poisoning was assessed using a repeated cross-sectional observational design. Participants were all residents in Spain aged 65 years or older from January 1, 2001, through December 31, 2022. Mortality was obtained from the General Mortality Register of the National Institute of Statistics.21 This register is based mainly on the Medical Death Certificate, although for deaths under judicial investigation, a Judicial Death Certificate is used.21,22 Forensic and toxicological information is added later during the coding and validation process.21 Drug poisoning deaths were defined as all deaths directly attributable to poisoning by psychoactive substances (excluding alcohol and tobacco), medicines, or other substances for therapeutic use, following the ICD-10.17 Thus, these deaths include all poisoning deaths by the aforementioned substances, irrespective of the pharmacological mechanism.14,23 Underlying causes of death were physician-certified and coded by the National Institute of Statistics following World Health Organization (WHO) rules.24 Drug poisoning deaths were classified by intent as accidental and intentional. Accidental poisonings (ICD-10 codes: X40-X44) occur when death is an unexpected consequence of consumption or exposure (ie, if the wrong drug is taken by mistake or inadvertently), excluding deaths caused by complications of drugs correctly prescribed and administered in appropriate therapeutic or prophylactic doses. Intentional poisonings (ICD-10 codes: X60-X64) occur when death results from purposeful self-inflicted poisoning or a suicide attempt. The ICD-10 recommends including as accidental deaths those self-inflicted poisonings in which it is not specified whether or not there was an intention to self-harm.25 The mentioned codes are internationally recommended for studies on fatal drug poisoning. Drug poisonings of undetermined intent (ICD-10 codes: Y10-Y14) and homicides (X85) were excluded from the analysis. Similarly, following WHO recommendations, mental and behavioral disorders due to psychoactive substance use (ICD-10 codes F11-F19) were also excluded, as the 2002 to 2003 ICD-10 update established a priority for X and Y codes over F codes in computing poisoning.26 In any case, the deaths excluded for this reason were very few, specifically 49 (Y10-Y14), 24 (X85), and 32 (F11-F19). Statistical Analyses Age-standardized mortality rates (ASMRs) per million person-years and their 95% CIs were obtained using the 2013 European reference population. Annual population counts by sex and age group obtained from the Spanish National Institute of Statistics were used as denominators for rate estimation. Time trends were characterized with Joinpoint regression, which models the trend of the natural logarithm of ASMR using connected linear segments, each with its own slope, which is transformed into annual percentage change (APC).27 The linear trend is downward or upward depending on whether the APC is greater or less than zero. Joinpoints represent the points in which a significant trend change occurs. A maximum of 4 joinpoints was specified, corresponding to the default recommendation of the Joinpoint Regression Program for time series of 22 to 26 observations.28 Model selection was performed using permutation tests (4499 permutations; 2-sided α = 0.05), with a minimum segment length of 3 observations at both ends and between joinpoints. Weighted least-squares log-linear regression was applied using weights equal to the inverse of the variance of the log-transformed ASMRs. Although the maximum number of permitted joinpoints was 4, all final models selected by the permutation procedure included no more than 1 Joinpoint, indicating that the data supported parsimonious models. The overall trend for the entire 2001 to 2022 period was characterized by the average APC, a weighted average of the APCs of the segments. Further, 95% CIs were obtained for the APCs and average APCs. When Joinpoint regression did not identify significant changes in trend but APC estimates were required for prespecified periods, weighted log-linear regression using the inverse of the variance of the log-transformed ASMRs was applied. Analyses were stratified separately by intent (accidental and intentional), sex (male and female), and age group (65-74, 75-84, and ≥85 years). Simultaneous stratifications by intent, sex, and age were not performed because of the small size of some subpopulations. The analyses were performed with using Stata BE 16.1 (StataCorp) and Joinpoint Regression Program, version 5.2.0.0 (National Cancer Institute). As a sensitivity analysis, Poisson regression models were fitted to age-specific death counts using the corresponding population as an offset to evaluate potential interaction effects between poisoning intent and sex and between poisoning intent and age group. Models included the corresponding main effects and interaction terms, and the statistical significance of these interactions was assessed using Wald tests. A second sensitivity analysis assessed the potential impact of alternative classifications of deaths coded as undetermined intent (ICD-10 codes Y10-Y14). All deaths of undetermined intent were alternatively reclassified as accidental poisonings and, separately, as intentional self-poisonings. ASMRs were recalculated under each scenario to evaluate the robustness of the main findings (Supplemental Table 3, available online at http://www.mcpiqojournal.org). Ethics The study protocol was approved by the Institutional Review Board of Universidad Complutense de Madrid (CE_20240912_08_SAL). The study was not prospectively registered. Results Distribution of Fatal Drug Poisonings Among Older People In Spain during 2001 to 2022, there were 3845 fatal drug poisonings among older adults, representing 19.5% of poisonings deaths at any age. Of these deaths, 59.1% occurred in women and 41.2% in people aged 85 years and older. Regarding the underlying cause of death, most fatal poisonings were classified as accidental (77.7%) and coded as poisoning by other and unspecified drugs, medicaments, and biological substances (ICD-10 codes; X44 and X64; 82.7%). The proportion of accidental poisoning was higher among aged 85 years or older (49.8%) and 75 to 84 years (34.6%) than among people aged 65 to 74 years (15.6%). However, intentional poisoning was higher among people aged 65 to 74 (57.1%) and 75 to 84 years (31.8%) than among people aged 85 years or older (11.1%) (Table 1). Among deaths occurring during 2016 to 2022 with a recorded ICD-10 T36 to T50 codes, the pharmacological groups most frequently identified were systemic and hematological agents (T45; n=634); agents primarily affecting the cardiovascular system (T46; n=356); diuretics and other unspecified drugs, medicaments, and biological substances (T50; n=102); and hormones and their synthetic substitutes and antagonists (T38; n=82) (Supplemental Table 4, available online at http://www.mcpiqojournal.org). Accidental poisonings accounted for most of T45 and T46 codes, whereas T50 codes represented a relatively larger share of intentional poisonings. Because T-code recording remained incomplete and nonuniform during the study period, these findings are presented for descriptive purposes only and should not be interpreted as representing the distribution of substances among all fatal drug poisonings.Table 1Number of Deaths and Percentages From Drug Poisoning Among People Aged ≥65 Years, by Sex, Age Group, and Intentionality: Spain, 2001-2022Characteristic2001-20152016-20222001-2022n%n%n%Total183020153845Sex Male73240.084041.7157240.9 Female109860.0117558.3227359.1Age (y) 65-7445624.949724.795324.8 75-8469337.961430.5130734.0 ≥8568137.290444.9158541.2Intent Accidentala143778.5155277.0298977.7 Intentionalb39321.546323.085622.3aAccidental poisonings: International Classification of Disease, 10th Revision, codes X40-X44.bIntentional poisonings: International Classification of Disease, 10th Revision, codes X60-X64. Table 1Number of Deaths and Percentages From Drug Poisoning Among People Aged ≥65 Years, by Sex, Age Group, and Intentionality: Spain, 2001-2022Characteristic2001-20152016-20222001-2022n%n%n%Total183020153845Sex Male73240.084041.7157240.9 Female109860.0117558.3227359.1Age (y) 65-7445624.949724.795324.8 75-8469337.961430.5130734.0 ≥8568137.290444.9158541.2Intent Accidentala143778.5155277.0298977.7 Intentionalb39321.546323.085622.3aAccidental poisonings: International Classification of Disease, 10th Revision, codes X40-X44.bIntentional poisonings: International Classification of Disease, 10th Revision, codes X60-X64. Number of Deaths and Percentages From Drug Poisoning Among People Aged ≥65 Years, by Sex, Age Group, and Intentionality: Spain, 2001-2022 Accidental poisonings: International Classification of Disease, 10th Revision, codes X40-X44. Intentional poisonings: International Classification of Disease, 10th Revision, codes X60-X64. During the entire period poisoning ASMRs per million person-years were quite similar in men (21.7) and women (19.9) and increased markedly with age, from 10.4 at age 65 to 74 years to 63.3 at age 85 years or older. Furthermore, ASMR from accidental poisoning (15.8) was 3 times higher than that from intentional poisoning (4.8) (Table 2).Table 2ASMRs and APCs From Drug Poisoning Among Persons Aged ≥65 Years, by Sex, Age Group, and Intentionality: Spain, 2001-2022aCharacteristic2001-20152016-20222001-2022ASMRs from drug poisoning per million person-yearsb (95%CI)Total15.5 (14.8-16.2)28.9 (27.7-30.2)20.6 (20.0-21.3)Sex Males16.1 (14.9-17.3)30.5 (28.5-32.6)21.7 (20.6-22.8) Females15.1 (14.2-16.0)27.8 (26.2-29.5)19.9 (19.1-20.8)Age (y) 65-747.7 (7.0-8.4)15.5 (14.2-17.0)10.4 (9.8-11.1) 75-8416.4 (15.2-17.7)29.1 (26.9-31.5)20.7 (19.6-21.8) ≥8546.7 (43.2-50.3)86.6 (81.0-92.4)63.3 (60.2-66.5)Intent Accidentalc12.0 (11.4-12.7)21.5 (20.5-22.6)15.8 (15.3-16.4) Intentionald3.4 (3.1-3.8)7.4 (6.7-8.1)4.8 (4.5-5.1)APC of mortality ratee (95%CI)Total16.0 (14.2-19.0)−3.5 (−6.8 to −0.8)9.1 (8.2-10.6)Sex Male14.1 (12.3-17.3)−5.3 (−10.8 to −1.4)8.2 (7.0-10.0) Female17.7 (14.9-23.3)−1.9 (−6.1 to 1.4)9.8 (8.4-12.3)Age (y) 65-74f8.4 (6.4-10.3)6.8 (2.4-10.8)7.3 (6.4-8.6) 75-8415.6 (13.3-20.2)−4.9 (−10.4 to 0.7)8.4 (7.0-10.6) ≥8524.7 (20.6-33.5)−10.2 (−15.7 to −5.8)11.8 (9.6-15.8)Intent Accidentalc19.6 (17.2-23.9)−7.4 (−11.3 to −4.2)9.9 (8.5-11.9) Intentionald,f5.9 (2.3-8.6)8.0 (3.2-12.8)7.3 (6.3-9.0)aAPC, annual percentage change; ASMR, age-standardized mortality rate; ICD-10, International Classification of Disease, 10th Revision.bRates were standardized by age using the direct method and the 2013 European Standard Population.cAccidental poisonings: ICD-10 codes X40-X44.dIntentional poisonings: ICD-10 codes X60-X64.eAPC represents the APC in the ASMR estimated using Joinpoint regression.fJoinpoint regression did not detect any significant trend change, but it was calculated for the periods 2001-2015 and 2016-2022 for comparative purposes using log-linear regression weighted by the inverse of the variance. Table 2ASMRs and APCs From Drug Poisoning Among Persons Aged ≥65 Years, by Sex, Age Group, and Intentionality: Spain, 2001-2022aCharacteristic2001-20152016-20222001-2022ASMRs from drug poisoning per million person-yearsb (95%CI)Total15.5 (14.8-16.2)28.9 (27.7-30.2)20.6 (20.0-21.3)Sex Males16.1 (14.9-17.3)30.5 (28.5-32.6)21.7 (20.6-22.8) Females15.1 (14.2-16.0)27.8 (26.2-29.5)19.9 (19.1-20.8)Age (y) 65-747.7 (7.0-8.4)15.5 (14.2-17.0)10.4 (9.8-11.1) 75-8416.4 (15.2-17.7)29.1 (26.9-31.5)20.7 (19.6-21.8) ≥8546.7 (43.2-50.3)86.6 (81.0-92.4)63.3 (60.2-66.5)Intent Accidentalc12.0 (11.4-12.7)21.5 (20.5-22.6)15.8 (15.3-16.4) Intentionald3.4 (3.1-3.8)7.4 (6.7-8.1)4.8 (4.5-5.1)APC of mortality ratee (95%CI)Total16.0 (14.2-19.0)−3.5 (−6.8 to −0.8)9.1 (8.2-10.6)Sex Male14.1 (12.3-17.3)−5.3 (−10.8 to −1.4)8.2 (7.0-10.0) Female17.7 (14.9-23.3)−1.9 (−6.1 to 1.4)9.8 (8.4-12.3)Age (y) 65-74f8.4 (6.4-10.3)6.8 (2.4-10.8)7.3 (6.4-8.6) 75-8415.6 (13.3-20.2)−4.9 (−10.4 to 0.7)8.4 (7.0-10.6) ≥8524.7 (20.6-33.5)−10.2 (−15.7 to −5.8)11.8 (9.6-15.8)Intent Accidentalc19.6 (17.2-23.9)−7.4 (−11.3 to −4.2)9.9 (8.5-11.9) Intentionald,f5.9 (2.3-8.6)8.0 (3.2-12.8)7.3 (6.3-9.0)aAPC, annual percentage change; ASMR, age-standardized mortality rate; ICD-10, International Classification of Disease, 10th Revision.bRates were standardized by age using the direct method and the 2013 European Standard Population.cAccidental poisonings: ICD-10 codes X40-X44.dIntentional poisonings: ICD-10 codes X60-X64.eAPC represents the APC in the ASMR estimated using Joinpoint regression.fJoinpoint regression did not detect any significant trend change, but it was calculated for the periods 2001-2015 and 2016-2022 for comparative purposes using log-linear regression weighted by the inverse of the variance. ASMRs and APCs From Drug Poisoning Among Persons Aged ≥65 Years, by Sex, Age Group, and Intentionality: Spain, 2001-2022a APC, annual percentage change; ASMR, age-standardized mortality rate; ICD-10, International Classification of Disease, 10th Revision. Rates were standardized by age using the direct method and the 2013 European Standard Population. Accidental poisonings: ICD-10 codes X40-X44. Intentional poisonings: ICD-10 codes X60-X64. APC represents the APC in the ASMR estimated using Joinpoint regression. Joinpoint regression did not detect any significant trend change, but it was calculated for the periods 2001-2015 and 2016-2022 for comparative purposes using log-linear regression weighted by the inverse of the variance. However, the poisoning distribution by intent varied according to age. Thus, at age 65 to 74 years, ASMRs from accidental and intentional poisonings were very similar (5.1 and 5.9, respectively), while at age 85 years or older, the first (59.5) was 16 times higher than the second (3.7). Furthermore, unlike ASMRs from accidental poisoning, which increases sharply with age, the ASMR from intentional drug poisoning decreases with age (Supplemental Table 2, available online at http://www.mcpiqojournal.org). Time Trends in Fatal Drug Poisonings Overall and by Substance Involved ASMR from poisoning, increased during 2001 to 2015 (slowly until 2007 and intensely thereafter) and decreased during 2015 to 2022, although the rates in 2022 continued to be much higher than those in 2001 (Figure 1). Regarding the substances involved, the group of nonspecific substances conditioned general trends, but a peak was seen in 2020 to 2022 in poisonings by the group of psychoactive medicines that includes tranquilizers, sleeping pills, antidepressants, or neuroleptics (X41 and X61) and the group of narcotics and psychodysleptics (X42 and X62) (Supplemental Figure 1, available online at http://www.mcpiqojournal.org). The Joinpoint regression identified a significant change in the temporal trend of poisoning rates in 2015, from a steep increase (APC, 16.0%) to a slight decrease (APC, −3.5%) (Figure 2, Table 2).Figure 1Age-standardized mortality rates (ASMRs) from drug poisoning per million person-years among persons aged ≥65 years, by sex, age group, and intentionality, in Spain, 2001-2022. ASMRs were standardized by age using the direct method and the 2013 European Standard Population. Accidental poisonings correspond to International Classification of Disease, 10th Revision (ICD-10) codes X40-X44 and intentional poisonings to ICD-10 codes X60-X64. Panels show trends overall (A) and stratified by sex, age group, and intent (B). Vertical dashed lines indicate key contextual events discussed in the article: implementation of the National Patient Safety Strategy/STOPP-START recommendations (2015) and onset of the COVID-19 pandemic (2020).Figure 2Observed (points) and Joinpoint regression-fitted lines age-standardized mortality rates (ASMRs) from drug poisoning per million person-years among persons aged ≥65 years in Spain, by sex, 2001-2022. Joinpoints indicate statistically significant changes in temporal trends. Rates were standardized by age using the direct method and the 2013 European Standard Population. Numbers displayed on the regression segments indicate the annual percentage change (APC) for the corresponding time interval. In which no Joinpoint was identified, a single APC for the entire study period (2001-2022) is shown. Corresponding 95% CIs are reported in Table 2. Age-standardized mortality rates (ASMRs) from drug poisoning per million person-years among persons aged ≥65 years, by sex, age group, and intentionality, in Spain, 2001-2022. ASMRs were standardized by age using the direct method and the 2013 European Standard Population. Accidental poisonings correspond to International Classification of Disease, 10th Revision (ICD-10) codes X40-X44 and intentional poisonings to ICD-10 codes X60-X64. Panels show trends overall (A) and stratified by sex, age group, and intent (B). Vertical dashed lines indicate key contextual events discussed in the article: implementation of the National Patient Safety Strategy/STOPP-START recommendations (2015) and onset of the COVID-19 pandemic (2020). Observed (points) and Joinpoint regression-fitted lines age-standardized mortality rates (ASMRs) from drug poisoning per million person-years among persons aged ≥65 years in Spain, by sex, 2001-2022. Joinpoints indicate statistically significant changes in temporal trends. Rates were standardized by age using the direct method and the 2013 European Standard Population. Numbers displayed on the regression segments indicate the annual percentage change (APC) for the corresponding time interval. In which no Joinpoint was identified, a single APC for the entire study period (2001-2022) is shown. Corresponding 95% CIs are reported in Table 2. Time Trends in Fatal Drug Poisonings by Sex and Age Group There were hardly any noticeable differences between sexes in the evolution of ASMRs (Figure 1). The Joinpoint model shows no significant differences between men and women, although women show a somewhat steeper upward slope during 2001 to 2015 and a gentler downward slope during 2015 to 2022 (Figure 2, Table 2). The mortality rate at age 65 and 74 years followed a similar upward trend throughout the study period, while at ages 75 to 84 and 85 years or older, it increased very sharply until approximately 2015 and then declined (Figure 1). The Joinpoint regression detects trend changes in 2015 for the previous 2 age groups, with more pronounced upward and downward trends at age 85 years or older than 75 to 84 years (Figure 3, Table 2).Figure 3Observed (points) and Joinpoint regression-fitted lines age-standardized mortality rates (ASMRs) from drug poisoning per million person-years among persons ≥65 years in Spain, by age group and intent, 2001-2022. Accidental poisonings correspond to International Classification of Disease, 10th Revision (ICD-10) codes X40-X44 and intentional poisonings to ICD-10 codes X60-X64. Rates were standardized by age using the direct method and the 2013 European Standard Population. Numbers displayed on the regression segments indicate the annual percentage change (APC) for the corresponding time interval. In which no Joinpoint was identified, a single APC for the entire study period (2001-2022) is shown. Corresponding 95% CIs are reported in Table 2. Observed (points) and Joinpoint regression-fitted lines age-standardized mortality rates (ASMRs) from drug poisoning per million person-years among persons ≥65 years in Spain, by age group and intent, 2001-2022. Accidental poisonings correspond to International Classification of Disease, 10th Revision (ICD-10) codes X40-X44 and intentional poisonings to ICD-10 codes X60-X64. Rates were standardized by age using the direct method and the 2013 European Standard Population. Numbers displayed on the regression segments indicate the annual percentage change (APC) for the corresponding time interval. In which no Joinpoint was identified, a single APC for the entire study period (2001-2022) is shown. Corresponding 95% CIs are reported in Table 2. Time Trends in Fatal Drug Poisonings by Intent ASMRs from accidental poisoning shows an evolution similar to overall poisoning, with a sharp increase until 2015 and a moderate decrease thereafter, whereas intentional poisoning shows a sustained moderate increase throughout the period (Figure 1). Joinpoint regression identified a linear trend change in ASMRs from accidental poisoning in 2015 from sharply upward (APC, 19.6%) to downward (APC, −7.4%). In contrast, no significant trend change was detected in intentional poisoning rate, with APCs of 5.9% and 7.4% during 2001 to 2015 and 2015 to 2022, respectively (Figure 3, Table 2). Sensitivity analyses using Poisson regression models confirmed significant interactions between poisoning intent and both sex (Wald χ2=7.91; P=.005) and age group (Wald χ2=460.85; P<.001) (Supplemental Table 5, available online at http://www.mcpiqojournal.org). These findings support the descriptive analyses by indicating that the association between poisoning intent and mortality differs according to both demographic characteristics, particularly across age groups. Discussion This study analyzed fatal drug poisonings among older adults in Spain, covering 2001 to 2022, using the basic cause of death from the general mortality register coded following ICD-10. There are only 2 previous studies in Spain on this topic. However, the first analyzes only poisonings at age 15 to 64 years,20 whereas the second focuses on people aged 65 years or older, without differentiating between age subgroups and ends in 2018.6 Our main results are discussed further. Fatal Drug Poisoning: A Different Phenomenon Among Younger and Older Adults During 2011 to 2022, the drug poisoning ASMR per million person-years among people aged 65 years or older (20.6) was very similar to that previously found among people aged younger than 65 years (22.6).20 This narrow inequality by age contrasts sharply with that identified in the specific Spanish register of mortality from acute reactions to psychoactive substances, on the basis of forensic and toxicological sources. Thus, during 2020 to 2022, the crude rate ratio from drug poisoning among people aged younger than 65 and 65 years or older was 6.3 in the aforementioned specific register compared with 1.1 in our study.29 This suggests that the differential involvement of psychoactive substances in drug poisonings before and after age 65 years could be the main reason for the differential inequality. Furthermore, although, in Spain, during the study period, there was hardly any multiple coding of substances using T codes, there is other indirect evidence that supports the aforementioned-hypothesis. Thus, during 2001 to 2022, the ASMRs from poisoning by narcotics and psychodysleptics (ICD-10 codes: X42 and X62) at ages younger than 65 and 65 years or older were 5.9% and 0.7%, respectively. If deaths from mental and behavioral disorders due to psychoactive substance use (ICD-10 codes F11-F19) were added, the drug poisoning ASMRs at ages younger than 65 and 65 years or older would have increased by 6.4% and 0.9%, respectively. Finally, population surveys show that in Spain older adults have a low prevalence of illicit drug use.30 All this evidence supports the hypothesis that, in Spain, during 2001 to 2022, drug poisoning mortality among younger and older adults responds to different causes. Although, among the former, fatal drug poisonings would be caused primarily by psychoactive substances, often illegal, among older adults, they would be caused primarily by prescription drugs, often involving multiple medications (polypharmacy), but rarely prescribed opioids. A high involvement of pharmaceuticals, including prescribed opioids, in drug poisoning deaths has also been found elsewhere.7,31, 32, 33 Polypharmacy has also been observed in nonfatal poisonings.34, 35, 36 In general, in high-income countries, older adults are a highly medicalized population.37 Furthermore, cognitive impairment and dependence on carers increase the likelihood of administration errors and nonadherence.38, 39, 40 Available reports from similar settings indicate that psychotropic medications (particularly benzodiazepines and antidepressants), analgesics, and certain cardiovascular drugs frequently appear in toxicological findings among older adults. Finally, concomitant alcohol consumption can enhance the toxicity of medications such as benzodiazepines.41,42 Differential Trends in Accidental and Intentional Fatal Drug Poisoning Among Older Adults In Spain, during 2001 to 2022, ASMR from drug poisoning among older adults increased globally, as also happened in other high-income countries.1,2,4 However, the time trends differed markedly by intent. Specifically, accidental poisonings reported a clear shift from an upward to a downward trend in 2015, whereas intentional poisonings exhibited a continuous increase. A previous study had already identified the upward trend, but not the 2015 trend change because the study ended in 2018.6 Several factors may have contributed to the upward trend during 2001 to 2015, especially the increasing coexistence of multiple chronic diseases requiring prescription of several drugs, which increases the risk of toxicity.6,31,43 Thus, the proportion of Spanish populations with simultaneous prescription of 5 drugs or more tripled during 2005 to 2015.44 Increases have also been documented elsewhere.37 The Joinpoint analysis identified an inflection in accidental drug poisoning mortality around 2015. This temporal coincidence with the implementation of several measures to improve the rational use of medicines introduced from 2015 onward is compatible with the hypothesis that these initiatives may have contributed to the subsequent decline in accidental poisoning mortality. However, the ecological design of this study does not permit causal attribution, and this temporal association should therefore be interpreted as hypothesis generating rather than causal. Specifically, these measures included the 2015 legal regulations that strengthened obligations regarding the prescription and dispensing of medicines and, particularly, the Patient Safety Strategy 2015-2020.45 This strategy aims to improve drug safety, implementing measures such as the following: (a) medication reconciliation during care transitions (ie, hospital admissions/discharges) to maintain treatment continuity and consistency; (b) protocols for high-risk medications (ie, anticoagulants), facilitating their identification and management through support tools; (c) detecting and reducing inappropriate medication (ie, benzodiazepines in some patients) and reviewing treatments in line with the Screening Tool of Older Persons' Prescriptions/Screening Tool to Alert to Right Treatment criteria46; (d) extension and interoperability between regions of electronic prescribing systems including clinical decision support; and (e) continuing education for health care professionals, patients, and caregivers on medication safety and treatment management. These measures are aligned with the WHO’s global challenge “Medication Without Harm” and are intended to reduce inappropriate prescribing, improving medication safety, and decrease medication-related adverse events, particularly among older adults receiving multiple medications.47 The upward trend in ASMRs from suicidal drug poisoning is consistent with the results of a study showing an increase in overall suicide rates among older people in recent years, highlighting the importance of considering demographic changes, variations in diagnostic coding, and other possible underlying trends in mental health.48 However, this finding should be interpreted with caution because several nonmutually exclusive mechanisms may contribute to the observed increase, and their relative contribution cannot be determined from mortality registry data alone. First, progressive improvements in death certification and forensic investigation during the study period may have reclassified some deaths previously coded as undetermined intent toward the intentional category, thereby contributing to an apparent upward trend whose magnitude cannot be quantified with the available data.22 Second, changes in prescribing practices and increasing availability of psychotropic medications, including benzodiazepines and sedative hypnotics, may have increased access to potentially lethal medications among older adults with suicidal intent.9 Finally, older adults have high psychosocial vulnerability characterized by a high prevalence of depression, loneliness, social isolation, bereavement, and chronic disease burden, all of which are well-established risk factors for suicidal behavior.49 These factors may also have contributed to the observed increase in intentional fatal poisonings, although this hypothesis cannot be assessed using mortality registry data alone. In the 65- to 74-year age group, in which the intentional poisoning rate is higher, greater autonomy in managing medication could facilitate access to lethal doses or combinations.9,49 Although mental health and suicide prevention programs are being implemented in Spain, the lack of interventions specifically tailored to older adults may limit their impact on intentional poisonings.50 Distinguishing the relative contribution of these mechanisms will require linkage of mortality data with individual-level clinical and prescribing information. Gender and Age Inequalities in Fatal Drug Poisoning Among Older Adults The small between-sex differences in poisoning rates among older people contrasts with the situation among younger people in which rate was 3 times higher in men than in women.20 This is probably due to the higher polypharmacy prevalence in older women and to the attenuation with age of pharmacokinetic/pharmacodynamic differences between sexes, which may reduce the excess mortality from drug poisoning found among young men associated with illicit drug use.7,51 There were also no substantial intersex differences in time trends, which are quite consistent with the absence of between-sex differences among older adults in population prevalence of polypharmacy or in its upward trend during 2005 to 2015.44 Some studies in high-income countries have found a stronger upward trend in drug poisoning mortality in men than in women.2,52 The results show a large increase in mortality risk from drug poisoning with age, which can be explained by the increased prevalence of multimorbidity, polypharmacy, and cognitive impairment with age,12 coupled with cumulative exposure to potentially toxic treatments and other risk factors, such as social isolation or lack of medication supervision. However, mortality from drug poisoning at age 75 years or older has declined since 2015, whereas this improvement was less evident among adults aged 65 to 74 years. The reasons for this different temporal pattern cannot be determined from the present study. In particular, our data do not include information on employment history, retirement status, or employment quality, precluding formal evaluation of these mechanisms. Adults aged 65 to 74 years occupy a transitional stage between working life and retirement and may therefore experience a distinct combination of occupational, economic, and psychosocial stressors. These factors are unlikely to be directly addressed by medication safety initiatives and may therefore represent an additional pathway contributing to drug poisoning mortality beyond medication safety. If these mechanisms contribute to the observed pattern, improving medication safety alone may be insufficient to further reduce fatal drug poisonings in this age group. Consistent with this hypothesis, Andreacchi et al53 reported substantially higher drug poisoning mortality among workers experiencing precarious employment, particularly among older workers. Although these findings derive from a Canadian working-age cohort and cannot be directly extrapolated to older adults in Spain, they suggest that labor-market vulnerability may represent an additional pathway influencing drug poisoning mortality beyond medication safety initiatives. Likewise, previous research in Spanish older adults suggests that financial and psychosocial circumstances may also influence health-related behaviors.54 Future studies integrating employment histories, socioeconomic information, prescribing data, and mortality records are needed to evaluate these mechanisms and to identify complementary prevention strategies for adults aged 65 to 74 years. The particularly high accidental poisoning mortality observed among adults aged 85 years or older, together with the pronounced post-2015 decline in this age group, also deserves specific consideration. Compared with younger older adults, individuals aged 85 years or older have a higher prevalence of cognitive impairment, dementia, multimorbidity, and functional dependency, conditions associated with difficulties in medication self-management and greater reliance on informal and formal caregivers for medication administration.38,39 Caregiver-related factors, including limited pharmacological training, increasingly complex multidrug regimens, inadequate medication reconciliation during transitions of care, and medication administration errors, may both increase baseline poisoning risk and represent potentially modifiable targets for intervention.40 Medication safety initiatives, including structured medication review approaches such as STOPP/START, are intended to address several of these potential mechanisms by identifying potentially inappropriate prescribing and promoting regular medication review in older adults.46 However, the present study cannot determine whether these mechanisms contributed to the favorable post-2015 trends observed in this age group, and this hypothesis would require linkage of mortality data with prescribing information and individual-level clinical data. Future studies should evaluate the contribution of cognitive impairment, caregiver involvement, and medication management to fatal drug poisoning among the oldest-old. Strengths and Limitations The main strengths of this study are the long study period, which allowed the evaluation of temporal trends over more than 2 decades, and the use of Joinpoint regression, which reliably identifies changes in drug poisoning mortality trends and facilitates the generation of hypotheses regarding potential determinants of these changes. Several limitations should also be acknowledged. First, the ICD-10 external-cause codes (X40-X44 and X60-X64) used in this study do not identify the specific substances responsible for death, do not allow distinction between single-agent and polypharmacy toxicity, and aggregate pharmacologically diverse exposures. Consequently, the observed temporal trends cannot be directly attributed to specific medication classes or underlying clinical mechanisms. Although ICD-10 T36 to T50 codes identifying the pharmacological substances involved became available from 2016 onward, their recording remained incomplete and nonuniform during the study period. Therefore, analyses by pharmacological class were limited to a descriptive summary of deaths with a recorded T-code (Supplemental Table 4, available online at http://www.mcpiqojournal.org) and should not be interpreted as representing the distribution of substances among all fatal drug poisonings. Consequently, reliable analyses of individual drug classes (eg, opioids, benzodiazepines or antidepressants) were not possible. Second, temporal improvements in death certification and coding practices may have influenced the observed trends independently of true changes in mortality. Previous studies have documented progressive improvements in forensic investigation and external-cause coding specificity in Spain during the study period. If these improvements increased the ascertainment of drug poisoning deaths over time, they could have contributed to the observed increase before 2015 and reduced the apparent magnitude of the subsequent decline. Consequently, any resulting bias would most likely be conservative with respect to our main findings. Furthermore, although the observed inflection in accidental poisoning mortality occurred around the same time as the implementation of medication safety initiatives in Spain, the ecological design of this study does not permit causal inference. Therefore, this temporal association should be interpreted as hypothesis generating rather than as evidence that these interventions caused the observed decline. Third, although deaths of undetermined intent (Y10-Y14) were excluded from the primary analyses, some degree of misclassification of intent cannot be ruled out, particularly among older adults. However, a bounding sensitivity analysis based on alternative reclassification of all Y10 to Y14 deaths as either accidental or intentional poisonings produced only minimal changes in ASMRs (Supplemental Table 3, available online at http://www.mcpiqojournal.org), suggesting that the observed divergence between accidental and intentional poisoning trends is robust to plausible alternative classifications of intent. Finally, mortality figures for 2020 should be interpreted with caution because the Coronavirus disease 2019 pandemic may have affected death registration processes. Moreover, given the annual resolution of the available data and the limited number of observations after 2020, the potential impact of the pandemic cannot be distinguished reliably from underlying secular trends. Conclusion In Spain, during 2001 to 2022, most fatal drug poisonings among older adults appear to have been caused by prescribed medications, likely multiple medications (polypharmacy). This would therefore be a rather different phenomenon from fatal drug poisonings in younger populations, in which illicit drugs are frequently involved. The time trend in drug poisoning mortality from accidental poisoning changed in 2015, from upward to downward, to which the safe medication programs implemented from that year onward may have contributed decisively. Despite this, the continued increase in fatal poisoning rates among people aged 65 to 74 years and in fatal intentional poisonings highlights the need to strengthen surveillance of poisonings and high-risk prescribed medications, the drug safety measures promoted since 2015,45, 46, 47 and suicide prevention strategies. Adults aged 85 years or older are the most vulnerable group, requiring close supervision of medication both at home and in residential settings. Potential Competing Interests The authors report competing interests. Ethics Statement The study protocol was approved by the Institutional Review Board of Universidad Complutense de Madrid (CE_20240912_08_SAL). The study was not prospectively registered.