Changing trends in drug overdose mortality in Spain, 2001–2022

Abstract

Aim

To measure the evolution of drug overdose mortality in Spain between 2001 and 2022.

Design, Setting, Participants

A repeated cross‐sectional observational study using nationwide mortality data from Spain, 2001–2022, among individuals aged 15–64 years.

Measurements

The outcome was overdose deaths [International Classification of Diseases, 10th revision (ICD‐10) codes: X40–X44, X60–X64, X85, Y10–Y14]. Annual age‐standardized mortality rates (ASMRs) per million person‐years were estimated. Joinpoint regression assessed changes in trends using annual percentage change (APC) and average APC (AAPC). Overdose deaths were grouped by the underlying drug categories involved: nonopioid analgesics, antipyretics and antirheumatics (X40/X60/Y10); antiepileptic, sedative‐hypnotic, antiparkinsonism and psychotropic drugs (X41/X61/Y11); narcotics and psychodysleptics (X42/X62/Y12); other drugs acting on the autonomic nervous system (X43/X63/Y13); other and unspecified drugs, medicaments and biological substances (X44/X64/Y14); and drugs, medicaments and biological substances in the context of assault (X85). Intentionality (based on ICD‐10 code characters) within the underlying drug categories was assessed. All analyses were stratified by sex.

Findings

In Spain, 15 862 overdose deaths occurred during 2001–2022. ASMR increased by 49.0% over the period, from 24.7 [95% confidence interval (95% CI) = 23.0, 26.5] to 36.8 (95% CI = 34.7, 39.0) deaths per million person‐years. APC showed a statistically significant change in trends in 2010, shifting from a −3.3% (95% CI = −7.3 to −1.7) decline to an increase of 3.4% (95% CI = −0.3 to 6.5). In 2018, a more abrupt change was observed, with the APC further increasing to 13.8% (95% CI = 9.3, 22.5). Females showed an earlier trend change (2006) with an AAPC 1.7 times higher (3.4%; 95% CI = 2.4, 4.7) than males (2.0%; 95% CI = 1.2, 2.7). Leading drug categories were other and unspecified drugs, medicaments and biological substances, followed by narcotics and psychodysleptics. In males, trends followed the overall pattern; in females, the antiepileptic, sedative‐hypnotic, antiparkinsonism and psychotropic drugs category predominated, mostly due to intentional overdoses.

Conclusions

Between 2001 and 2022, drug overdose deaths in Spain increased substantially, beginning in 2010 and further accelerated in 2018, though population rates remain low. The most prevalent underlying ICD‐10 drug categories were other and unspecified drugs, medicaments and biological substances—likely reflecting polydrug use—and narcotics and psychodysleptics. The observed rise in deaths involving antiepileptic, sedative‐hypnotic, antiparkinsonism and psychotropic drugs, particularly intentional overdoses among females, highlights the need for sex‐based interventions.

INTRODUCTION

In Europe, overdose deaths have been historically dominated by heroin, followed by cocaine‐related fatalities, although the population of people who inject drugs (PWID) is declining and aging [1, 2, 3, 4, 5, 6]. This trend may partly reflect the long‐term impact of harm reduction services, opioid substitution therapies and the treatment of injection‐related infections, which have likely contributed to moderating mortality [7, 8, 9]. However, in recent years, most overdose deaths show polydrug toxicity, with opioids – typically in combination with other substances – remaining the most frequent drug implicated in overdoses [6]. Additionally, stimulant‐related deaths are likely to be under‐reported and appear more commonly in younger age groups [6]. Furthermore, the proportion of drug‐induced deaths involving benzodiazepines has increased in countries like Denmark and Spain [6]. These shifts suggest limitations in current interventions, primarily designed to address heroin use, which may be insufficient to respond to ongoing patterns of drug use.

Still, the available data indicate that the scale and characteristics of overdose mortality in Europe differ substantially from those in the USA, where overdose deaths are a major contributor to premature mortality and are primarily driven by illicit synthetic opioids such as fentanyl [10, 11, 12, 13]. According to the European Union Drugs Agency (EUDA), more than 6400 drug‐induced deaths were reported in the European Union (EU) in 2022 among individuals aged 15–64 years, corresponding to a mortality rate of 22.5 deaths per million population [6], approximately 15 times lower than that reported in the US (326 per million in 2022) [14]. Nonetheless, rising trends have been observed in several European countries, including Ireland, Sweden and the UK, where overdose mortality rates exceed 50 deaths per million population [6]. These disparities are likely driven by differences in drug availability, patterns in drug use, and access to harm reduction and treatment services [13].

In Spain, while overdose mortality remains low (1266 deaths in 2022) compared with other European countries, such as Germany (1631 deaths in 2022) [6, 15], changing trends in prescription drugs use, particularly opioids and hypnosedatives, raise concern because of their potential to increase drug‐related harms. Evidence from other high‐income countries has shown that increases in the prescribing of these drug classes have preceded rises in drug‐related mortality [16, 17, 18, 19, 20].

Data from the Spanish Agency of Medicines and Health Products (AEMPS) show that opioid prescriptions have more than doubled, from 9.9 defined daily doses per 1000 inhabitants per day (DHD) in 2010 to 22.3 in 2022 [21]. This increase has been primarily driven by tramadol and fentanyl prescribing [21, 22]. The number of individuals receiving at least one opioid prescription rose from 335 379 to 722 838 between 2010 and 2018, and the total opioid dosage per capita – measured in morphine milligram equivalents (MME/c) – has nearly tripled. Fentanyl MME/c more than tripled, accounting for 34.4% of the total MME/c by 2018 [23]. This trend is consistent with previous findings showing a more than 10‐fold increase in fentanyl use and a pronounced, greater than sevenfold, rise in tramadol use between 2007 and 2019 [24].

These patterns reflect broader trends observed in other European countries, including Denmark, France, Germany and the UK, where prescription opioid use has also increased – though generally at lower levels than those seen in the USA or Canada [13, 25]. Additionally, in 2021, Spain was the second‐highest consumer of prescription anxiolytics in Europe, after Portugal, and ranked seventh for hypnosedatives [26]. Use increased from 83.0 to 90.5 DHD between 2010 and 2023 [27].

Overdose deaths represent missed opportunities for intervention. Monitoring consumption patterns, identifying at‐risk populations and tracking trends are essential for effective harm reduction strategies. In Spain, however, recent data on overdose deaths remain limited. This study aimed to assess nationwide trends in drug overdose mortality from 2001 to 2022 among individuals aged 15–64 years.

METHODS

Design, study population, data sources and variables

The evolution of overdose mortality was assessed using a repeated cross‐sectional observational study design.

Participants were all residents of Spain aged 15–64 years between 2001 and 2022. Mortality data came from annual deidentified death records from the Spanish National Statistics Institute (INE), which extracts a single underlying cause of death for each deceased person (the disease that started the chain of events leading to death) from medical death certificates following the recommendations of the World Health Organization, and codes them using the International Classification of Diseases, 10th Revision (ICD‐10) [28]. The annual population figures were also obtained from the INE. The national mortality data and population figures represent the entire resident population of Spain (for further details on mortality and population data sources, see Table S1).

The primary outcome was drug overdose deaths, as suggested by prior literature [29, 30, 31, 32], and included all deaths with any of the following underlying causes of death, irrespective of intentionality: X40–X44, X60–X64, X85 and Y10–Y14 (for details on the individual ICD‐10 codes included, see Table S1). These deaths were first analyzed in aggregate, and then categorized, grouped by ICD‐10 codes referring to the same substance types, regardless of intent classification. These groupings are referred to throughout the text as ‘underlying drug categories’. The categories defined were: nonopioid analgesics, antipyretics and antirheumatics (X40, X60 and Y10); antiepileptic, sedative‐hypnotic, antiparkinsonism and psychotropic drugs (X41, X61 and Y11); narcotics and psychodysleptics (X42, X62 and Y12); other drugs acting on the autonomic nervous system (X43, X63 and Y13); other and unspecified drugs, medicaments and biological substances (X44, X64 and Y14); and drugs, medicaments and biological substances in the context of assault (X85). Finally, overdose deaths – both overall and within each underlying drug category – were further classified by intent: unintentional/undetermined (X4x/Y1x) versus intentional (X6x).

The negative control outcome was external causes of death (V01–Y98, excluding drug overdose codes) [33].

We included the following covariates: calendar years between 2001 and 2022, sex (males/females, as recorded in the INE) and age (in 5‐year groups and broader categories: 15–29, 30–49 and 50–64 years).

Statistical analysis

We described the characteristics of drug overdose deaths using absolute numbers and proportions.

Annual overdose age‐standardized mortality rates (ASMRs) per million person‐years and their corresponding 95% confidence intervals (95% CIs) were calculated using weights from the 2013 European standard population and plotted using the RStudio geom_smooth() function. By default, geom_smooth() uses LOESS (locally estimated scatterplot smoothing), a non‐parametric regression method that fits multiple local polynomial regressions to the data. ASMRs were also plotted by broad age groups (15–29, 30–49 and 50–64 years).

Joinpoint regression models of ASMRs were used to identify significant changes in mortality trends over the study period. Joinpoint software fitted weighted least‐squares regression models to the ASMRs to estimate the annual percentage change (APC) and the average annual percentage change (AAPC) [34], which were then plotted. Analyses were set to allow a maximum of four joinpoints across the period [35]. The permutation tests for model (number of joinpoints) significance were set at an overall alpha level of 0.05 [36].

Finally, to assess variations in overdose mortality across underlying drug categories, annual ASMRs were calculated and plotted for each drug category. Similarly, rates by intentionality categories were calculated and plotted over time.

All analyses were stratified by sex.

Data analysis was performed using Stata BE 16.1, R 4.4.0/RStudio and Joinpoint 5.3.0. These statistical analysis programs have been obtained from the licenses of the authors' institutions.

The Carlos III Health Institute review board approved the study proposal and waived the requirement for informed consent. Ethics committee approval was not required. The analysis was not pre‐registered and the results should be considered exploratory.

RESULTS

Sample characteristics

Over the study period (2001–2022), overdose deaths totaled 15 862, representing 1.2% of all deaths (n = 1 323 313) recorded in individuals aged 15–64 years. The majority of deaths occurred in males (74.7%) (Table 1), with rates up to 3.8 times higher than in females. The age group of 30–49 years accounted for 9983 deaths (62.9%). Most overdose deaths were of unintentional or undetermined intent (71.4%; 69.9% unintentional and 1.5% undetermined), with intentional overdoses representing 28.6%.

TABLE 1.

Number of deaths and age‐standardized mortality rate (ASMR) from drug overdose^a per million person‐years among people aged 15–64 years by age group, calendar year and sex. Spain, 2001–2022.

	Overall			Males				Females
	n	ASMR	95% CI	n	%	ASMR	95% CI	n	%	ASMR	95% CI
Total	15 862	22.6	22.2, 22.9	11 842	74.7	33.1	32.5, 33.7	4020	25.3	11.8	11.4, 12.2
Age, years
15–29	2074	11.6	11.1, 12.1	1633	78.7	17.8	16.9, 18.7	441	21.3	5	4.6, 5.5
30–49	9983	31.5	30.8, 32.1	7792	78.1	48.3	47.3, 49.4	2191	21.9	14	13.4, 14.6
50–64	3805	20.6	20.0, 21.3	2417	63.5	26.7	25.6, 27.8	1388	36.5	14.8	14.0, 15.6
Year
2001	767	24.7	23.0, 26.5	617	80.4	38.9	35.8, 42.1	150	19.6	10.1	8.6, 11.9
2002	695	22.2	20.6, 24.0	559	80.4	34.7	31.9, 37.8	136	19.6	9.3	7.8, 11.1
2003	664	20.9	19.4, 22.6	523	78.8	31.8	29.1, 34.7	141	21.2	9.7	8.1, 11.4
2004	678	21.3	19.7, 23.0	530	78.2	32.3	29.6, 35.2	148	21.8	10	8.4, 11.8
2005	648	19.6	18.1, 21.2	511	78.9	29.9	27.4, 32.7	137	21.1	8.8	7.4, 10.5
2006	569	17.2	15.8, 18.7	451	79.3	26.6	24.2, 29.2	118	20.7	7.5	6.2, 9.0
2007	685	20.6	19.1, 22.3	542	79.1	31.7	29.1, 34.6	143	20.9	9.1	7.6, 10.7
2008	675	20.1	18.6, 21.6	522	77.3	30.2	27.6, 32.9	153	22.7	9.5	8.0, 11.1
2009	601	17.7	16.3, 19.2	444	73.9	25.3	23.0, 27.8	157	26.1	9.7	8.3, 11.4
2010	569	16.9	15.5, 18.3	424	74.5	24.4	22.1, 26.9	145	25.5	9	7.6, 10.6
2011	527	15.8	14.5, 17.2	385	73.1	22.5	20.3, 24.9	142	26.9	8.9	7.5, 10.5
2012	582	17.3	15.9, 18.7	430	73.9	24.9	22.6, 27.4	152	26.1	9.4	7.9, 11.0
2013	640	19.2	17.7, 20.7	443	69.2	26.1	23.7, 28.7	197	30.8	12	10.4, 13.8
2014	723	22.1	20.5, 23.8	525	72.6	31.5	28.9, 34.4	198	27.4	12.4	10.7, 14.3
2015	605	18.5	17.1, 20.1	419	69.3	25.3	22.9, 27.8	186	30.7	11.6	10.0, 13.5
2016	705	21.3	19.8, 23.0	521	73.9	31.8	29.1, 34.7	184	26.1	11.1	9.5, 12.8
2017	712	22.0	20.4, 23.7	497	69.8	30.3	27.7, 33.1	215	30.2	13.6	11.8, 15.6
2018	673	20.7	19.2, 22.4	496	73.7	30.5	27.9, 33.4	177	26.3	10.8	9.3, 12.6
2019	823	25.4	23.7, 27.2	595	72.3	36.7	33.8, 39.8	228	27.7	14.1	12.3, 16.0
2020	1055	31.9	30.0, 33.9	773	73.3	46.4	43.1, 49.8	282	26.7	17	15.1, 19.1
2021	1061	32.4	30.5, 34.5	770	72.6	47.2	44.0, 50.7	291	27.4	17.5	15.5, 19.6
2022	1205	36.8	34.7, 39.0	865	71.8	52.8	49.3, 56.5	340	28.2	20.7	18.6, 23.1

^aCodes X40–X44, X60–X64, X85 and Y10–Y14 of the International Classification of Diseases, 10th Revision. n, number of cases; %, percentage; ASMR, age‐standardized mortality rate per million person‐years; 95% CI, 95% confidence interval.

Trends in drug overdose ASMRs

Between 2001 and 2022, overdose ASMR increased by 49.0% over the study period, from 24.7 (95% CI = 23.0, 26.5) to 36.8 (95% CI = 34.7, 39.0) deaths per million person‐years (Table 1). The evolution of overdose ASMRs (Figure 1) showed that after an initial decline, there was a clear upward trend starting from 2010, reaching peak levels at the end of the study period.

FIGURE 1.

Evolution of annual age‐standardized mortality rate (ASMR) from drug overdose¹ per million person‐years among people aged 15–64 years. Spain, 2001–2022. ¹Codes X40–X44, X60–X64, X85 and Y10–Y14 of the international classification of diseases, 10th revision. ASMR, age‐standardized mortality rate per million person‐years. The figure was made in RStudio using the geom_smooth() function. By default, geom_smooth() uses LOESS (locally estimated scatterplot smoothing), a non‐parametric regression method that fits multiple local polynomial regressions to the data. Each point represents the ASMR per million person‐years. The blue line represents the smoothed trend in drug overdose ASMR per million person‐years over time. Gray shading represents the confidence interval around the trend line, indicating the variability in the trend estimation.

By age category, the group aged 30–49 years had the highest ASMRs, which declined steadily until around 2010, and then plateaued before rising sharply again after 2015. In contrast, the group aged 50–64 years remained relatively stable until 2010, but a sustained increase began thereafter, converging with the rates of the former group by 2022 (Figure S1).

By intentionality (Figure S2; Table S2), both unintentional/undetermined intent and intentional overdoses showed an upward trend beginning in 2010. However, the unintentional/undetermined intent category exhibited rates that were 1.5–4.3 times higher than the intentional overdose category across the study period.

Assessment of changes in overdose death trends using joinpoint analyses

Joinpoint analysis of overall overdose ASMRs identified a significant change in trends in 2010, followed by a more abrupt change in 2018 (Figure 2a). The ASMR slope measured by the APC shifted from a decline of −3.3% (95% CI = –7.3 to −1.7) to an increase of 3.4% (95% CI = –0.3 to 6.5) in 2010, and further increased to 13.8% (95% CI = 9.3, 22.5) in 2018, until the end of the period (Table 2). Although not significant, females showed an earlier shift towards an upward overdose mortality trend compared with males (2006 and 2011, respectively) (Figure 2b). Both sexes experienced a significant upward trend in 2018, consistent with the overall trend. Despite the difference in the magnitude of rates, females exhibited an AAPC that was 1.7 times higher than that for males (2.0%; 95% CI = 1.2, 2.7), with an average annual increment of 3.4% (95% CI = 2.4, 4.7) (Table 2).

FIGURE 2.

Joinpoint regression analysis of annual age‐standardized mortality rate (ASMR) from drug overdose¹ per million person‐years among people aged 15–64 years: (a) overall and (b) by sex. Spain, 2001–2022. ¹Codes X40–X44, X60–X64, X85 and Y10–Y14 of the international classification of diseases, 10th revision. The figures were generated using Joinpoint regression program software with a maximum of four joinpoints. Each square represents the observed age‐standardized mortality rate (ASMR) per million person‐years. The lines in each colored segment represent different trends over time, and the point where two colors meet represents joinpoints indicating shifts in these trends. Segments are always represented by the same color throughout plots: segment 1, blue; segment 2, green; segment 3, red.

TABLE 2.

Joinpoint regression annual percentage change (APC) and average annual percentage change (AAPC) in age‐standardized mortality rate (ASMR) from drug overdose ^a per million person‐years, by sex. Spain, 2001–2022.

	APC					AAPC
	Segment	Period	Value (%)	95% CI	P	Value (%)	95% CI	P
Overall	1	2001–2010	−3.3*	−7.3 to −1.7	0.01	2.3*	1.7, 2.9	0.00
	2	2010–2018	3.4	−0.3 to 6.5	0.07
	3	2018–2022	13.8*	9.3, 22.5	0.00
Males	1	2001–2011	−3.9*	−8.8 to −0.3	0.05	2.0*	1.2, 2.7	0.00
	2	2011–2018	4.4	−6.6 to 8.0	0.26
	3	2018–2022	13.6*	7.8, 23.9	0.00
Females	1	2001–2006	−3.4	−15.3 to 3.3	0.21	3.4*	2.4, 4.7	0.00
	2	2006–2018	3.4	−2.8 to 7.5	0.08
	3	2018–2022	12.9*	6.1, 25.1	0.01

^a Codes X40–X44, X60–X64, X85 and Y10–Y14 of the International Classification of Diseases, 10th Revision.

For the joinpoint regression, a maximum of four joinpoints has been set. %, percentage; 95% CI, 95% confidence interval; P, probability.

^*Annual percentage change (APC) or the average annual percentage change (AAPC) significantly differs from zero at the alpha = 0.05 level.

Drug overdose causes of death by underlying categories

Figure 3 displays overdose deaths by underlying drug category. Most deaths during the entire study period were coded under other and unspecified drugs, medicaments and biological substances (X44, X64 and Y14) and narcotics and psychodysleptics (X42, X62 and Y12). While narcotics and psychodysleptics were the most prevalent at the start of the period (X42 X62 and Y12), this decreased substantially as other and unspecified drugs, medicaments and biological substances increased (X44, X64 and Y14), becoming the leading cause from 2004 onwards. The ASMR of other and unspecified drugs, medicaments and biological substances (X44, X64 and Y14) peaked in 2019, while narcotics and psychodysleptics (X42, X62 and Y12) increased steadily starting in 2019, equaling the former category in 2020 and maintaining a relatively parallel increasing trend since then. Interestingly, antiepileptic, sedative‐hypnotic, antiparkinsonism and psychotropic drugs (X41, X61 and Y11) also experienced an upward trend starting in 2019, similar to narcotics and psychodysleptics (X42, X62 and Y12); however, this peaked in 2020 at a lower ASMR, and the rate then remained at that value.

FIGURE 3.

Evolution of annual age‐standardized mortality rate (ASMR) from drug overdose¹ per million person‐years among people aged 15–64 years, by underlying drug category. Spain, 2001–2022. ¹Codes X40–X44, X60–X64, X85 and Y10–Y14 of the international classification of diseases, 10th revision. ASMR, age‐standardized mortality rate per million person‐years. The figure was made in RStudio using the geom_smooth() function. By default, geom_smooth() uses LOESS (locally estimated scatterplot smoothing), a non‐parametric regression method that fits multiple local polynomial regressions to the data. Each point represents the ASMR per million person‐years. Each line represents the smoothed trend in drug overdose ASMR per million person‐years over time. Drug categories are presented using descriptive names to represent ICD‐10 codes included in each group, to facilitate interpretation. Drug categories: grey – nonopioid analgesics, antipyretics and antirheumatics (X40, X60, Y10); red – antiepileptic, sedative‐hypnotic, antiparkinsonism and psychotropic drugs (X41, X61 and Y11); green – narcotics and psychodysleptics (X42, X62 and Y12); orange – other drugs acting on the autonomic nervous system (X43, X63 and Y13); light blue – other and unspecified drugs, medicaments and biological substances (X44, X64 and Y14); brown – drugs, medicaments and biological substances in the context of assault (X85). For further code definitions, see Table S1.

Differences in overdose ASMR by underlying drug category groups and sex are shown in Figure S3. While males followed the overall trend, females had higher rates in the antiepileptic, sedative‐hypnotic, antiparkinsonism and psychotropic drugs category (X41, X61 and Y11). Figure S4 further reveals that for females, intentional overdoses (X61) predominated within this category, whereas unintentional overdoses (X41) were more common in males. Additionally, females in the other and unspecified drugs, medicaments and biological substances (X44, X64 and Y14) group showed higher rates in the intentional category (X64), continuing until nearly the end of the study period.

Negative control

The negative control outcome showed no similar changes in the annual evolution of ASMRs (Figure 4) or joinpoint regression analysis (data not shown) during the study period.

FIGURE 4.

Evolution of annual age‐standardized mortality rate (ASMR) from external causes other than those related to drug overdose¹ per million person‐years among people aged 15–64 years. Spain, 2001–2022. ¹Codes V01–Y98 (excluding drug overdose X40–X44, X60–X64, X85 and Y10–Y14 codes) of the international classification of diseases, 10th revision. ASMR, age‐standardized mortality rate per million person‐years. The figure was made in RStudio using the geom_smooth() function. By default, geom_smooth() uses LOESS (locally estimated scatterplot smoothing), a non‐parametric regression method that fits multiple local polynomial regressions to the data. Each point represents the ASMR per million person‐years. The blue line represents the smoothed trend in external causes ASMR per million person‐years over time. Gray shading represents the confidence interval around the trend line, indicating the variability in the trend estimation.

DISCUSSION

Overdose mortality in Spain among individuals aged 15–64 years increased by 49.0% between 2001 and 2022, primarily owing to overdoses with unintentional/undetermined intent. The overdose mortality rate declined during the first decade but reversed in 2010 and then accelerated after 2018. Joinpoint analysis confirmed significant trend‐change points at both these time points. Among females, the increase began earlier (in 2006) and was steeper, with an AAPC that was 1.7 times higher than that recorded in males. The contribution of each underlying drug category varied over time, although the most relevant contributors were other and unspecified drugs, medicaments and biological substances (X44, X64 and Y14), followed by narcotics and psychodysleptics (X42, X62 and Y12). These categories reached similar levels in 2020 and have continued to rise in parallel. By sex, males followed the overall pattern, while females showed higher rates in the antiepileptic, sedative‐hypnotic, antiparkinsonism and psychotropic drugs category (X41, X61 and Y11), where intentional overdoses predominated.

Increasing trends in overdose mortality in recent years: how does Spain compare with other countries?

The upward trend we observed in Spain mirrors patterns observed in other countries. In England and Wales, ASMRs increased from 33.9 in 2010 to 53.9 per million person‐years in 2022 [37]; in the USA, ASMRs increased from 123.0 to 326.0 during the same period [14]. Although Spain’s overdose ASMR remains lower (rising from 16.9 in 2010 to 36.8 in 2022), the recent acceleration is concerning. In the USA, more recent rises have been linked to synthetic opioids, particularly illicit fentanyl [14, 38]. In Europe, this dynamic has been less prominent; however, Estonia and the UK have reported deaths resulting from illicitly manufactured opioids and nitazenes [39, 40, 41, 42]. In Spain, while illicit synthetic opioids do not appear to be driving overdose deaths, the recent rise in mortality highlights the need to understand which substances are implicated. In this context, increases in opioid prescribing have been observed and could play a role, potentially alongside other emerging substance use patterns [24, 43].

Understanding overdose drug involvement: insights from substance use treatment admissions

The most frequent drug categories contributing to overdose deaths were other and unspecified drugs (X44, X64 and Y14), likely reflecting polydrug use as per coding recommendations, which assign these codes when substances involved belong to different underlying drug categories (e.g. cocaine, coded as X42, X62 and Y12, and benzodiazepines, coded as X41, X61 and Y11) [44]. This was followed by narcotics and psychodysleptics (X42, X62 and Y12), likely encompassing opioids and cocaine [28], and antiepileptic, sedative‐hypnotic, antiparkinsonism and psychotropic drugs (X41, X61 and Y11).

Substance use treatment admissions are a key source of information on consumption patterns. In 2022, of 45 873 admissions, cocaine was the leading substance (47.6%, n = 21 818), followed by cannabis (27.4%) and opioids (18.7%). Polydrug use was reported in approximately 40.0% of admissions [15]. Heroin‐related admissions have declined since 2004, while cocaine‐related admissions peaked in 2008 and 2019 [15]. Although admissions for tramadol, synthetic cannabinoids and fentanyl are low (approx. 200 total in 2022) [15], first‐time admissions for opioids other than heroin have increased, suggesting a possible role of prescription opioid misuse [45].

For sedative‐hypnotic and psychotropic drugs (X41, X61 and Y11) [28, 44], hypnosedative‐related admissions increased by 43.7% from 2018 (n = 662) to 2022 (n = 951) [15].

Injecting drug use, a major risk factor for overdose deaths [46, 47], has declined by 79.9% in new admissions since 2001 [15]. Between 2015 and 2022, the total number of first‐time injectors decreased slightly, while the proportion reporting recent injections (<12 months) increased from 35.4% to 45.8% [15]. However, this increase alone is unlikely to account for the rise in overdose deaths.

Contextualizing overdoses with population survey data

In the Survey on Alcohol and Other Drugs (EDADES) of people aged 15–64 years, the past‐month prevalence of use of opioid analgesics was approximately 4%, while lifetime use was approximately 14.0% [15]. Among lifetime users, codeine and tramadol predominated, while the reported lifetime use of fentanyl has increased 7.4‐fold between 2018 and 2022, reaching approximately 14% of lifetime‐opioid users [48]. This suggests considerable exposure, likely through medical prescribing, which may contribute to diversion or misuse. These trends align with prior findings in Spain of increased prescriptions for opioids, especially fentanyl and tramadol [23, 24, 49].

Past‐month hypnosedative use in the general population aged 15–64 years (EDADES) rose from 3.7% in 2005 to 9.7% in 2022 (APC = 4.7%), especially among females aged 55–64 years, where prevalence reached 21.4% [50].

Insights from judicially investigated deaths

Toxicology data on judicially investigated deaths caused by acute reactions to psychoactive substances, reported by the Spanish Observatory of Drugs and Addictions (OEDA), show an upward trend in overdoses since 2013, and especially after 2018, aligning with our results [15]. Fatalities most often involve multiple substances, and single‐substance cases are rare. Cocaine has been among the most frequently detected substances in overdose deaths, being the second most frequent drug reported in 2022, detected in 59.9% of deaths (n = 531), often in combination with opioids (50.7%). Opioids were present in 51.2% of deaths, and methadone was the most frequently detected opioid (25.8%), despite few treatment admissions, suggesting diversion. Detections of fentanyl (n = 33) and tramadol (n = 104) have increased more than threefold and twofold, respectively, since 2018 [15].

Hypnosedatives have become the most frequently detected drug class since 2014 (usually with polydrug use), particularly among females, and in suicide‐related deaths in both sexes [15].

Judicially investigated deaths show a shift in median age from <40 years to >40 years after 2010 [15]. This aligns with the rising ASMRs reported in the age group of 50–64 years, and coincides temporally with the overall reversal in mortality trend. This demographic change likely reflects an aging population of people who use drugs (PWUD). Age‐related factors – such as comorbidity burden, reduced drug metabolism and impaired self‐monitoring – may increase overdose risk [51]. The predominance of deaths in this group coded as X44, X64 or Y14 (data not shown) suggests a growing impact of polysubstance use in this population. This demographic shift is likely an additional contributor to the rising burden of overdose deaths.

Sex differences in overdose mortality

While overdose deaths remain more common in males [6, 15], our data show a more rapid increase in females, with an AAPC that is 1.7 times higher than that in males. Similar trends have been observed in Scotland [52]. Hypothesized contributors include increased drug use among ‘hidden’ populations, a higher burden of physical and mental health conditions, aging among females who use drugs, reduced access to harm reduction services, shifts in drug use patterns (e.g. polysubstance use) and changes in caregiving or relationship dynamics [52].

Prescription drug use may also play a contributory role. Hypnosedatives and prescription opioids are more commonly prescribed to females – particularly older females – and this is also reflected in population‐based surveys and the increasing detection of these substances in toxicology reports in female fatal overdoses [15, 26, 50, 53]. Among first‐time treatment admissions for opioid use, females show a higher proportion of admissions involving opioids other than heroin [45]. Furthermore, we observed intentional overdoses – especially involving sedative‐hypnotic and psychotropic drugs – were more common in females, suggesting distinct risk profiles and pathways leading to overdose.

These findings highlight the importance of developing sex‐specific strategies, including prescription monitoring, expanded access to mental health care and improved engagement of females in substance use treatment services.

Impact of the COVID‐19 pandemic

The COVID‐19 pandemic occurred during the ongoing acceleration in overdose deaths that began in 2018. Mandatory stay‐at‐home orders implemented in March 2020 in Spain intermittently disrupted access to harm reduction services through restrictions to movement, the temporary closure of some treatment centers and reduced operating hours [48, 54]. Treatment admissions dropped in 2020, with heroin‐related admissions reaching an all‐time low, and had not returned to pre‐pandemic levels by 2021 [48]. Additional factors – such as social isolation, the sudden loss of routine, income uncertainty, temporary employment suspensions, unemployment and increased stress – may have unintentionally increased overdose risks. These disruptions likely created high‐risk use environments, marked by interruptions in treatment continuity, reduced service access, unstable drug supplies and reduced tolerance after a period of abstinence [54, 55, 56]. Notably, prescriptions for benzodiazepines, other psychiatric medications and opioids increased during this period in Spain, mostly in females [21, 26, 27, 50, 57, 58].

Drivers of mortality increase

The rising involvement of multiple drug categories, prescribing trends and toxicology findings collectively suggest that polydrug use – particularly involving hypnosedatives, cocaine and opioids – has become a major driver of overdose mortality in Spain in recent years. The strong and sustained presence of cocaine in both treatment admissions and forensic toxicology supports its central role. Methadone, the most frequently detected opioid in forensic data, despite low treatment admission numbers, points to possible diversion. The detection of other opioids, alongside increasing first‐time treatment admissions for non‐heroin opioids, suggests a growing role for prescription opioids, which is supported by the marked rise in opioid prescribing over the past decade [22]. Hypnosedatives have emerged as key contributors to overdose deaths, especially in suicides and among females, consistent with rising prescription rates in older females. These patterns coincide with a demographic shift in overdose mortality. The marked increase in ASMRs among those aged 50–64 years suggest an aging cohort of PWUD who are likely at increased vulnerability for overdose, with polysubstance‐related deaths as a main driver. While a modest increase in recent injection drug use among first‐time treatment admissions has been observed between 2015 and 2022, its contribution seems limited. These observations highlight the importance of targeting polydrug use patterns – especially those involving hypnosedatives, cocaine and opioids – in future prevention strategies [59].

Strengths and limitations

This study leverages high‐quality nationwide overdose mortality data across a 22‐year period [60]. However, contributing cause‐of‐death codes relevant to overdoses (ICD‐10: T36–T50) are under‐recorded in Spain. Since their implementation in 2016, only 6.4% of overdose deaths have included these T‐codes, limiting any insight into the substances involved. As these codes are more frequently recorded in judicialized deaths, they are unlikely to represent the full population of overdose cases. Consequently, T‐codes were not analyzed, and complementary data sources were used to better understand drug consumption patterns in Spain. Improving the completeness and consistency of toxicology data in mortality statistics, as has been done in other European countries, would enhance surveillance [44].

Mortality registration may also suffer from under‐reporting and misclassification [61]. Cross‐validation for 2022 suggests that overdose deaths in Spain may be under‐reported by as much as 20% [6], implying that the actual number of deaths is underestimated.

In line with the WHO recommendations, we excluded ICD‐10 F‐codes (F11–F19), which represent mental and behavioral disorders arising from psychoactive substance use, rather than acute overdose events [44]. Overdose deaths are classified under external cause codes – X‐codes for unintentional and intentional poisonings and Y‐codes for poisonings of undetermined intent – to capture the event leading to death rather than chronic conditions. During our study period, F11–F19 codes (excluding F17, tobacco) accounted for 1265 deaths, 69.2% of which occurred before 2010, with the largest proportion occurring before 2006.

Our study period overlaps with significant events, such as the 2008 economic crisis and the COVID‐19 pandemic. Both events may have influenced overdose mortality trends through effects on health service access, socio‐economic conditions or the completeness and quality of death certification. However, prior evidence suggests that the 2008 crisis did not substantially alter overdose mortality trends in Spain [62, 63].

CONCLUSION

Overdose mortality in Spain has increased substantially since 2010, with a marked acceleration after 2018. While rates remain comparatively lower than in other high‐income countries, the recent rise underscores the need for enhanced surveillance and complete toxicological reporting to better characterize the substances involved to inform targeted interventions. Overdose deaths were most frequently attributed to categories of other and undetermined drugs, narcotics and psychodysleptics and sedative‐hypnotic and psychotropic drugs, patterns consistent with widespread polydrug use and an increasing role of prescription medications. National toxicological data further support a growing involvement of hypnosedatives, cocaine and prescription opioids. These patterns are unfolding alongside a demographic shift in overdose deaths, with a rising burden among adults aged ≥50 years, likely reflecting an aging cohort of PWUD with increased vulnerability. These findings highlight the need for improved monitoring of prescribing practices, greater clinical awareness of overdose risks associated with polydrug use and expanded access to prevention and treatment services, and age‐specific interventions directed to the aging population who use drugs. Finally, given the steeper increases in intentional overdose mortality observed among females, prevention efforts should prioritize prescription drug‐related risks – particularly involving hypnosedatives – and address the unique clinical and social factors affecting females who use these medications.

AUTHOR CONTRIBUTIONS

Florencia Giné: Data curation (equal); formal analysis (equal); funding acquisition (equal); investigation (equal); software (equal); writing—original draft (equal); writing—review and editing (equal). Marta Donat: Data curation (equal); formal analysis (equal); funding acquisition (equal); investigation (equal); software (equal); writing—original draft (equal); writing—review and editing (equal). Juan Miguel Guerras: Investigation (equal); methodology (equal); supervision (equal); writing—original draft (equal); writing—review and editing (equal). Gregorio Barrio: Funding acquisition (equal); investigation (equal); methodology (equal); project administration (equal); supervision (equal); writing—original draft (equal); writing—review and editing (equal). María José Belza: Investigation (equal); methodology (equal); supervision (equal); writing—original draft (equal); writing—review and editing (equal). Julieta Politi: Conceptualization (equal); data curation (equal); formal analysis (equal); funding acquisition (equal); investigation (equal); methodology (equal); software (equal); supervision (equal); writing—original draft (equal); writing—review and editing (equal). Enrique Regidor: Conceptualization (equal); investigation (equal); methodology (equal); project administration (equal); resources (equal); supervision (equal).

DECLARATION OF INTERESTS

None to declare. The findings and conclusions in this report are those of the authors and do not necessarily represent the official position of Carlos III Health Institute or the other institutions with which the authors are affiliated.

Supporting information

Table S1 Mortality data according to the International Classification of Diseases, 10th Revision (ICD‐10) codes, with corresponding descriptions and population data used in this article.

Table S2 Number of cases and age‐standardized mortality rate (ASMR) from drug overdose1 per million person‐years among people aged 15 to 64 years, by intentionality. Spain, 2001‐2022.

Figure S1 Evolution of annual age‐standardized mortality rate (ASMR) from drug overdose1 per million person‐years among people aged 15 to 29, 30 to 49 and 50 to 64 years. Spain, 2001‐2022.

Figure S2 Evolution of annual age‐standardized mortality rate (ASMR) from drug overdose1 per million person‐years among people aged 15 to 64 years, by intentionality. Spain, 2001‐2022.

Figure S3 Evolution of annual age‐standardized mortality rate (ASMR) from drug overdose1 per million person‐years among people aged 15 to 64 years, by underlying drug category and sex. Spain, 2001‐2022.

Figure S4 Evolution of annual age‐standardized mortality rate (ASMR) from drug overdose1 per million person‐years among people aged 15 to 64 years, by underlying drug category, intentionality and sex: a) antiepileptic, sedative‐hypnotic, antiparkinsonism and psychotropic drugs (X41/X61/Y11), b) narcotics and psychodysleptics (X42/X62/Y12), and c) other and unspecified drugs, medicaments and biological substances (X44/ X64/Y14). Spain, 2001‐2022.

Giné F, Donat M, Guerras JM, Barrio G, Belza MJ, Politi J, et al. Changing trends in drug overdose mortality in Spain, 2001–2022. Addiction. 2026;121(2):261–272. 10.1111/add.70121

DATA AVAILABILITY STATEMENT

The datasets analyzed in this work can be requested from the National Institute of Statistics of Spain.