National trends in hospitalisation, in-hospital mortality and resource use for SLE in Spain from 2016 to 2022

Abstract

Objective

SLE is a complex autoimmune disease that causes significant health issues. In this study, SLE-related healthcare episodes, in-hospital mortality and resource use in Spain from 2016 to 2022 were assessed.

Methods

In this retrospective, nationwide, repeated cross-sectional study, International Classification of Diseases, 10th Revision, data from Spain’s National Hospital Discharge Registry were used to identify healthcare episodes with SLE (or its subtypes) coded as the principal diagnosis. Temporal trends were analysed and group comparisons were made using χ², Fisher’s exact test, Student’s t test and Mantel-Haenszel trend test.

Results

A total of 24 151 SLE episodes were identified, with the majority occurring in women (86.9%) and individuals aged 15–65 years (88.2%). SLE episodes increased from 3002 (6.5/100 000 population) in 2016 to 4226 (8.8/100 000 population) in 2022 (p<0.001). Lupus nephritis episodes rose from 525 (1.13/100 000 population) to 956 (2.0/100 000 population; p<0.001). Relative to all other hospitalisation episodes, SLE episodes had longer hospital stays (8.7 vs 7.1 days; p<0.001) and intensive care unit (ICU) stays (8.4 vs 4.9 days; p=0.031). In-hospital mortality was low and stable. In unadjusted comparisons, mortality was lower among SLE episodes than among other hospitalisation episodes (0.3% vs 1.2%; p<0.001). A larger proportion of in-hospital deaths during SLE episodes than during all other hospitalisation episodes occurred in individuals aged 15–65 years (42.5% vs 15.1%; p<0.001). Although more deaths occurred in women, mortality was higher among men in unadjusted comparisons (0.7% vs 0.3%; p<0.001).

Conclusions

SLE-related healthcare episodes increased in Spain from 2016 to 2022, but in-hospital mortality remained stable. SLE episodes involved longer hospital and ICU stays than other episodes. These results reflect healthcare utilisation patterns, not disease severity or patient outcomes.

WHAT IS ALREADY KNOWN ON THIS TOPIC

• Previous Spanish registry-based studies have described hospitalisation and mortality patterns in SLE, but nationally representative data for the International Classification of Diseases, 10th Revision era and recent healthcare activity are limited.

WHAT THIS STUDY ADDS

• Between 2016 and 2022, healthcare episodes with SLE coded as the principal diagnosis increased in Spain, while episode-level in-hospital mortality remained low and stable. These episodes had longer hospital and intensive care unit stays than other hospital episodes in unadjusted descriptive comparisons.

HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

• These findings provide updated information on hospital activity and resource utilisation that may support healthcare planning and future epidemiological research, but should not be interpreted as patient-level estimates of disease burden, severity or prognosis.

Introduction

SLE is a complex chronic autoimmune disease characterised by the loss of immunological tolerance and multisystem involvement. Its clinical manifestations range from mild constitutional and mucocutaneous symptoms to severe organ-threatening complications, particularly affecting the kidneys, central nervous system, lungs and cardiovascular system. SLE follows a relapsing-remitting course and is associated with substantial morbidity, cumulative organ damage and excess mortality relative to the general population.^1–4 Although population-based studies suggest that a long-term decline in SLE-related mortality has occurred, this trend has been heterogeneous across demographic groups and healthcare settings.⁵

SLE affects predominantly women of reproductive age, with a female: male ratio of 9:1–15:1.^{1 6 7} Its incidence and prevalence vary widely by geographical region and ethnicity.⁸ The heaviest burden has been reported among African-American and Hispanic women in North America,^9–11 whereas lower rates have been reported for European countries such as the UK and for Taiwan, Australia and Denmark.^12–16 Despite marked improvements in survival, resulting in 5-year survival rates exceeding 90% in many cohorts, standardised mortality ratios remain two to five times higher for patients with SLE than for the general population, due mainly to cardiovascular, infectious and renal complications.^{9 10 16 17}

Administrative databases enable the analysis of large population-level datasets and long-term trends in SLE-related hospitalisation.¹⁸ The findings of such analyses, however, should be interpreted as patterns of healthcare utilisation, rather than direct measures of disease activity or patient-level outcomes, due to coding inaccuracies, limited clinical detail and the use of hospital episodes rather than individual patients. In Spain, almost 100 000 hospital episodes with SLE coded as the principal diagnosis were recorded between 1997 and 2015, with admissions for active SLE declining but those in which SLE was coded as a comorbidity, associated mainly with cardiovascular disease, infections and malignancies, increasing.^{16 19} The EPISER 2016 (study of the prevalence of rheumatic diseases in adult population in Spain) indicated that the estimated prevalence of SLE was about 210 cases per 100 000 adult inhabitants of Spain,²⁰ and national mortality data from 1980 to 2018 show a sustained decline since 1999.²¹ In this context, the minimum basic dataset (MBDS) of the Spanish National Hospital Discharge Registry provides a valuable framework for the updating of nationwide data on hospital episodes with SLE coded as the principal diagnosis and in-hospital outcomes.²² In this study, we examined specialised healthcare episodes with SLE coded as the principal diagnosis in Spain from 2016 to 2022, with analysis of their epidemiological characteristics, temporal trends, in-hospital mortality and healthcare resource utilisation.

Methods

Study design

This retrospective, nationwide, repeated cross-sectional study was performed with data from all healthcare episodes for which SLE was coded as the principal diagnosis in Spain from 2016 to 2022. Four clinically relevant SLE-associated diagnoses were also examined: drug-induced lupus erythematosus (DILE; International Classification of Diseases, 10th Revision (ICD-10) code M32.0), lupus nephritis (LN; codes M32.14 and M32.15), lupus-associated endocarditis (LSE; code M32.11) and lupus-associated pericarditis (LP; code M32.12). Full ICD-10 code lists and classification rules are provided in online supplemental table S1.

Data sources

Data were obtained from the MBDS compiled by Spain’s Ministry of Health, which contains aggregated, non-linkable records of emergency and conventional hospital admissions, as many centres’ emergency departments have short-stay and medium-stay units.²² During the study period, the coverage of the Registro de Actividad de Atención Sanitaria Especializada (RAE-CMBD) expanded progressively as additional healthcare activity settings were incorporated. According to the ministry of health user manual, the coverage of several specialised care modalities increased gradually over time, while emergency department activity became available in the public statistical portal from 2018 onwards. These changes should be considered when interpreting temporal trends in both overall registry activity and disease-specific healthcare episodes. The unit of analysis was the healthcare episode recorded in the MBDS rather than the individual patient. Consequently, repeated episodes from the same individual could not be identified or linked over time, and all analyses should be interpreted at the episode level rather than the patient level. Population data from the national public health surveillance system were used to calculate annual hospitalisation rates per 100 000 inhabitants. National population figures and mortality data were obtained from the National Institute of Statistics.²³

Variables

Data on the following variables were collected: number of episodes, age group, sex, hospital size (by bed count), in-hospital mortality, mortality stratified by age and sex, regional hospitalisation rates, costs, hospital length of stay (LOS) and intensive care unit (ICU) LOS. In-hospital mortality was defined as death recorded in the discharge status of the healthcare episode. ICU LOS was derived from the RAE-CMBD variable ‘Mean ICU Length of Stay’ and was calculated only for hospitalisation episodes requiring ICU admission.^{22 23}

Statistical analysis

Comparisons were performed between SLE episodes and all other non-SLE hospital episodes. In addition, each SLE-associated manifestation (DILE, LN, LSE and LP) was compared both with non-SLE hospital episodes and with the overall group of SLE episodes.

Categorical variables were compared using the χ² test or Fisher’s exact test, as appropriate. Between-group differences (including those between sexes) were described using ORs with 95% CIs, with no adjustment for potential confounders. Temporal trends in hospitalisation frequency, mortality and rates were assessed using the Mantel-Haenszel test.

Continuous variables (LOSs and costs) were analysed using aggregated annual mean values (one observation per calendar year). Results are presented as means with SD across the seven annual observations, and comparisons between groups were performed using independent Student’s t-test applied to these annual mean values. Given the ecological nature of the data and the limited number of timepoints, these analyses were used as descriptive comparative measures.

Costs were estimated using standardised diagnosis-related group-based methods.^{24 25} The study used the All Patient Refined Diagnosis Related Groups (APR-DRG)-based mean cost estimates provided directly by the RAE-CMBD database; therefore, costs were analysed as reported and were not adjusted for inflation.

Small discrepancies in sex-stratified and age-stratified totals were attributable to data suppression for anonymisation and confidentiality purposes in the national registry. Suppressed or non-disclosed records were not imputed and were excluded from the corresponding stratified analyses and percentage calculations.

No formal sample size calculation was performed because all eligible healthcare episodes recorded in the national RAE-CMBD during the study period were included in the analyses.

All analyses were two sided, with p<0.05 considered to be significant and were performed using SPSS V.25.0 (IBM Corporation, Armonk, New York, USA). No multivariable model was applied, as the analyses were designed as episode-level descriptive comparisons using aggregated national administrative data.

Patient and public involvement

No patient or public involvement was required.

Results

Sample

From 2016 to 2022, a total of 132 437 484 specialised healthcare episodes were recorded in Spain’s MBDS, with SLE coded as the principal diagnosis in 24 151 (0.02%) of cases (table 1). Of the SLE-related episodes, DILE accounted for 61 (0.25%), LN accounted for 4998 (20.7%), LSE accounted for 76 (0.31%) and LP accounted for 404 (1.67%; online supplemental table S2).

Table 1. Demographic and clinical characteristics of SLE-related hospitalisation episodes versus all episodes, 2016–2022.

Category	Total episodes		Total SLE		SLE vs total episodes
	n	%	n	%	OR	CI (95%)	P value
No. of episodes	132 437 484	100.0	24 151	100.0
Age <15 years	17 024 903	12.9	556	2.3	0.16	0.15 to 0.17	<0.001*
Age 15–65 years	72 995 342	55.1	21 288	88.2	6.06	5.82 to 6.30	<0.001*
Age >65 years	42 417 239	32.0	2307	9.5	0.22	0.21 to 0.23	<0.001*
Male	62 144 349	46.9	3169	13.1	0.17	0.16 to 0.18	<0.001*
Female	70 232 371	53.0	20 982	86.9	5.86	5.65 to 6.09	<0.001*
<200 beds	37 371 376	28.2	2722	11.3	0.32	0.31 to 0.33	<0.001*
200–500 beds	37 926 852	28.6	6170	25.6	0.86	0.83 to 0.88	<0.001*
501–1000 beds	34 376 211	26.0	9147	37.9	1.74	1.69 to 1.78	<0.001*
>1000 beds	22 763 045	17.2	6112	25.3	1.63	1.59 to 1.68	<0.001*
In-hospital mortality	1 528 870	1.2	80	0.3	0.28	0.23 to 0.35	<0.001*

Cumulative numbers and percentages of episodes are shown.

^*p<0.05.

Temporal trends

Rates of healthcare episodes with SLE coded as the principal diagnosis increased over time. Expressed per 100 000 population, the crude annual rate increased from 6.5 in 2016 to 8.8 in 2022 (p<0.001). The rate for LN also increased from 1.1 in 2016 to 2.0 in 2022 (p<0.001). All rates remained below 10 per 100 000 population throughout the study period (table 2).

Table 2. Temporal trends in hospitalisation episodes in Spain (2016–2022).

Category	2016		2017		2018		2019		2020		2021		2022		P value
	n	Rate	n	Rate	n	Rate	n	Rate	n	Rate	n	Rate	n	Rate
Total Spanish population	46 427 100		46 510 461		46 715 383		47 087 778		47 344 852		47 424 595		47 940 295
Total episodes	8 175 427	17 604	8 655 697	18 609	18 577 688	39 787	20 141 755	42 787	19 562 847	41 320	26 558 122	56 003	30 765 948	64 181	<0.001*
SLE	3002	6.5	2959	6.4	3149	6.7	3511	7.5	3451	7.3	3853	8.1	4226	8.8	<0.001*
DILE	6	0.01	11	0.02	9	0.02	10	0.02	7	0.01	6	0.01	12	0.03	0.76
LN	525	1.1	515	1.1	656	1.4	748	1.6	740	1.6	858	1.8	956	2.0	<0.001*
LSE	6	0.01	6	0.01	16	0.03	12	0.03	10	0.02	12	0.03	14	0.03	0.11
LP	46	0.09	46	0.09	60	0.1	50	0.1	62	0.1	67	0.1	73	0.2	0.004*

Annual numbers and crude hospitalisation rates per 100 000 inhabitants are shown. Crude rates were calculated using the mid-year (July 1) population from the Spanish National Institute of Statistics. Temporal trends were assessed using the Mantel-Haenszel χ² test.

^*p<0.05.

DILE, drug-induced SLE; LN, lupus nephritis in SLE; LP, pericarditis associated with SLE; LSE, endocarditis associated with SLE.

Demographic characteristics

Marked female predominance was observed, with 86.9% of SLE episodes occurring in women. Most SLE-related admissions were of individuals aged 15–65 years, and the proportion of these admissions was larger for SLE episodes than for all hospital episodes (88.2% vs 55.1%; OR 6.06, 95% CI 5.82 to 6.30; p<0.001; table 1). This age distribution was observed across most subgroups (online supplemental table S2).

Hospital characteristics

SLE-related episodes were more frequent in large hospitals, with the largest proportion occurring in centres with 501–1000 beds (37.9% vs 26.0% of all hospitalisation episodes; OR 1.74, 95% CI 1.69 to 1.78; p<0.001). Hospitals with >1000 beds also handled a larger proportion of SLE episodes (25.3% vs 17.2% of all hospitalisation episodes; OR 1.63, 95% CI 1.59 to 1.68; p<0.001). A similar pattern was observed across SLE subtypes. LN episodes were most frequent in hospitals with 501–1000 beds (42.8% vs 26.0% of all episodes; OR 2.13, 95% CI 2.01 to 2.25; p<0.001), whereas a larger proportion of LP episodes was recorded in hospitals with >1000 beds (29.5% vs 17.2% of all episodes; OR 2.01, 95% CI 1.62 to 2.49; p<0.001).

Lengths of stay

The mean hospital stay was longer for SLE episodes than for all other hospitalisation episodes (8.71 vs 7.13 days; p<0.001). Among SLE-associated manifestations, DILE showed the longest LOS (17.55 days; p=0.04), whereas LN showed the shortest LOS (7.04 days), with no significant difference compared with all other hospitalisation episodes (p=0.80) (online supplemental table S4).

The mean ICU stay was longer for SLE episodes than for all hospitalisation episodes (8.4 vs 4.9 days; p=0.031). Similarly, the mean ICU stay for LN was longer (6.8 vs 4.9 days; p=0.041). In contrast, ICU stays for DILE (1.6 vs 4.9 days; p=0.017) and LP (1.9 vs 4.9 days; p<0.001) were shorter (online supplemental table S5).

Costs

The mean cost per episode was lower for SLE episodes than for all hospitalisation episodes (€2373.42 vs €2898.72; p<0.001). Among SLE subtypes, costs were higher for LSE (€6358.35; p=0.002) and LP (€4239.05; p=0.035; online supplemental table S6).

Mean ICU costs for SLE episodes did not differ from those for all hospitalisation episodes (€13 676.16 and €11 660.48, respectively; p=0.15). Among SLE subtypes, ICU costs did not differ for LN (€12 477.27; p=0.61) or LSE (€9437.49; p=0.40), and were lower for DILE (€1952.33; p<0.001) and LP (€7446.06; p<0.001; online supplemental table S7).

Mortality

In-hospital deaths were recorded for 80 (0.3%) of the 24 151 SLE episodes. This proportion was lower than that for all hospitalisation episodes (1.2%; OR 0.28, 95% CI 0.23 to 0.35; p<0.001; tables 1 and 3). In-hospital mortality was also lower for LN episodes than for all other hospitalisation episodes (0.5%; OR 0.41, 95% CI 0.27 to 0.62; p<0.001), whereas DILE and LP showed no significant differences. No death was recorded for the LSE episodes (online supplemental table S2). No significant difference in mortality was observed between SLE subgroups and SLE (table 4).

Table 3. Age and sex distributions of deaths during SLE-related and all hospitalisation episodes, 2016–2022.

Category	Total episodes		SLE		SLE vs total episodes
	n	%	n	%	OR	CI (95%)	P value
Mortality	1 528 870	100	80	100	0.28	0.23 to 0.35	<0.001*
Age <15 years	8853	0.6	1	1.3	2.17	0.30 to 15.62	0.43
Age 15–65 years	231 228	15.1	34	42.5	4.15	2.66 to 6.46	<0.001*
Age >65 years	1 288 789	84.3	45	56.2	0.24	0.15 to 0.37	<0.001*
Male	818 948	53.6	23	28.8	0.35	0.22 to 0.57	<0.001*
Female	709 557	46.4	57	71.2	2.86	1.76 to 4.64	<0.001*

Cumulative numbers and percentages of episodes are shown. Percentages represent the distribution of in-hospital deaths over the entire study period within each category, rather than annual mortality rates or case-fatality proportions. For each category, ORs were calculated by comparing SLE episodes with all other hospital episodes within the same category, which served as the reference group.

^*p<0.05.

Table 4. Age and sex distributions of deaths for SLE subgroups versus all SLE episodes, 2016–2022.

Category	SLE total episodes		DILE		DILE vs SLE total episodes			LN		LN vs SLE total episodes			LSE		LSE vs SLE total episodes			LP		LP vs SLE total episodes
	n	%	n	%	OR	CI (95%)	P value	n	%	OR	CI (95%)	P value	n	%	OR	CI (95%)	P value	n	%	OR	CI (95%)	P value
Mortality	80	100.0	1	100.0	5.07	0.69 to 37	0.08	24	100.0	1.57	0.97 to 2.56	0.06	0	0.0	—	—	—	2	100.0	1.51	0.37 to 6.17	0.56
Age <15 years	1	1.3	0	0.0	—	—	—	0	0.0	—	—	—	0	0.0	—	—	—	0	0.0	—	—	—
Age 15–65 years	34	42.5	0	0.0	—	—	—	12	50.0	1.55	0.59 to 4.05	0.38	0	0.0	—	—	—	1	50.0	1.44	0.09 to 23.84	0.80
Age >65 years	45	56.2	1	100.0	2.39	0.09 to 60.55	0.38	12	50.0	0.70	0.27 to 1.82	0.46	0	0.0	—	—	—	1	50.0	0.73	0.04 to 12.16	0.83
Male	23	28.8	0	0.0	—	—	—	5	20.8	0.56	0.18 to 1.73	0.31	0	0.0	—	—	—	1	50.0	2.55	0.15 to 42.51	0.49
Female	57	71.2	1	100.0	1.25	0.05 to 31.75	0.53	19	79.2	1.80	0.58 to 5.59	0.31	0	0.0	—	—	—	1	50.0	0.39	0.02 to 6.56	0.50

Cumulative numbers and percentages of in-hospital deaths for the 2016–2022 study period are shown. For each category, ORs were calculated by comparing each SLE subgroup with the remaining SLE episodes within the same category, which served as reference group.

—, not estimated due to zero events; DILE, drug-induced SLE; LN, lupus nephritis in SLE; LP, pericarditis associated with SLE; LSE, endocarditis associated with SLE.

Mortality differed between men and women for all hospitalisation episodes (1.3% vs 1.01%; OR 0.76, 95% CI 0.76 to 0.77; p<0.001) and SLE episodes (0.7% vs 0.3%; OR 0.37, 95% CI 0.23 to 0.61; p<0.001). For LN episodes, mortality proportions were similar between sexes (0.6% and 0.5%, respectively; OR 0.76, 95% CI 0.28 to 2.04; p=0.58; online supplemental table S8).

Women accounted for a larger proportion of deaths during SLE episodes than during all hospitalisation episodes (71.2% vs 46.4%; OR 2.86, 95% CI 1.76 to 4.64; p<0.001), with a similar pattern observed for LN (79.2% vs 46.4%; OR 4.39, 95% CI 1.64 to 11.75; p<0.001). Accordingly, men accounted for a smaller proportion of deaths during SLE episodes than during all hospitalisation episodes (28.8% vs 53.6%; OR 0.35, 95% CI 0.22 to 0.57; p<0.001), and the pattern was similar for LN (20.8% vs 53.6%; OR 0.23, 95% CI 0.09 to 0.61; p<0.001; table 3 and online supplemental table S3).

Deaths during SLE episodes were less concentrated among older patients than were deaths during all hospitalisation episodes: patients aged >65 years accounted for 56.2% of SLE deaths and 84.3% of all in-hospital deaths (OR 0.24, 95% CI 0.15 to 0.37; p<0.001). Conversely, the 15–65 years old age group accounted for a larger proportion of SLE deaths (42.5% vs 15.1%; OR 4.15, 95% CI 2.66 to 6.46; p<0.001), and deaths were infrequent among patients aged <15 years (1.3% and 0.6%, respectively; OR 2.17, 95% CI 0.30 to 15.62; p=0.43; table 3 and online supplemental table S3).

Population-based rate of in-hospital deaths during SLE episodes increased significantly for all hospital episodes during the study period (p<0.001), whereas no significant temporal trends were observed for SLE or LN episodes (online supplemental table S9).

Geographical distribution

Rates varied geographically. Madrid had the highest hospitalisation rate (17.8 per 100 000 inhabitants), followed by Andalusia (11.5 per 100 000 inhabitants) and Catalonia (6.3 per 100 000 inhabitants). Rates were lowest for Spain’s northern regions, including Asturias, Cantabria, the Basque Country, Navarra and La Rioja (figure 1). The geographical distribution of hospitalisations for LN and other SLE subgroups is shown in online supplemental figure S1.

Figure 1. Average annual SLE hospitalisation rates (per 100 000 population) by region in Spain (2016–2022).

Discussion

This nationwide study updates registry-based research on hospitalisations for SLE in Spain.^{16 26} Population-based rates of in-hospital deaths remained low and stable between 2016 and 2022, despite a slight increase in SLE-related hospitalisation rates and a significant rise in overall hospital in-hospital mortality during the same period. Because the analysis was restricted to healthcare episodes in which SLE was coded as the principal diagnosis, these findings likely reflect patterns of hospital activity rather than patient-level outcomes or prognosis, consistent with recent hospital-based studies.^27–30

In-hospital mortality rates were lower for SLE episodes than for the overall hospitalised population, although this finding should be interpreted cautiously, given the heterogeneity of the comparator group in terms of age, comorbidities and admission characteristics. These differences may reflect case mix rather than an intrinsic prognostic advantage.^31–33

In contrast to observations of higher mortality for LN than for other SLE subtypes in long-term outcome studies, likely related to cumulative organ damage and recurrent flares,^{5 34} we observed no such pattern in this episode-based analysis. Mortality proportions were numerically highest for DILE (1.6%), followed by LN and LP (both 0.5%), compared with 0.3% for SLE overall, although no significant differences were observed between subgroups. These findings should be interpreted cautiously due to the lack of patient-level follow-up and the small numbers of events in some categories.^{16 34 35}

Sex differences were also observed; although more deaths occurred in women due to their predominance among admissions, mortality rates were higher in men. These results are derived from unadjusted comparisons and may reflect differences in age, comorbidities and case mix, rather than independent prognostic effects. Episodes involving individuals aged 15–65 years accounted for a greater proportion of deaths in SLE than in the overall hospital population, reflecting differences in age distribution rather than age-specific risk, as reported previously for European cohorts.^{35 36}

Although the study period included the coronavirus disease 2019 (COVID-19) pandemic, no increase in lupus-related in-hospital mortality was observed during these years. Mortality remained stable across SLE and all analysed manifestations. Evidence suggests that the impact of COVID-19 on SLE was often indirect, through the disruption of access to care and treatment delays, whereas severe outcomes and mortality were heterogeneous and relatively uncommon.^{37 38} Nonetheless, hospital episode-based analyses may lead to the underestimation of the broader patient-level consequences of the pandemic.

Our findings differ from the previous observation of an increase in the in-hospital mortality associated with SLE in Spain, driven largely by infections, cardiovascular disease and malignancies.¹⁶ This discrepancy may be explained in part by methodological differences, as our analysis was restricted to admissions with SLE coded as the principal diagnosis, whereas previous studies also included cases in which SLE was recorded as a comorbidity.^{16 26 33} However, principal diagnosis coding does not necessarily identify the underlying cause of death and may be influenced by coding practices.

Hospitalisation patterns in this study differed from those reported previously. Earlier Spanish registry studies documented a progressive increase in total SLE-related admissions, together with a decline in admissions interpreted as involving active SLE.¹⁶ In our study, the absolute number of admissions increased over time, with a parallel increase in population-based rates and overall hospital activity during the study period. LN showed significant increases in absolute number and population-based rates. These differences may reflect changes in case definition, coding practices and the transition from ICD-9 to ICD-10, which entails greater diagnostic granularity.^{16 33 39} Therefore, the observed increase in hospital episodes with SLE coded as the principal diagnosis should not necessarily be interpreted as evidence of a true increase in disease occurrence, as changes in coding practices, diagnostic ascertainment and healthcare organisation may also have contributed to this trend. Unlike previous nationwide analyses, the present analysis involved the stratification of SLE admissions by subtype.

SLE accounted for only a small proportion (0.02%) of overall hospital activity, consistent with results from administrative studies conducted in other settings.^{31–33 40–42} These proportions remain lower than estimates derived from population-based studies, likely because hospital discharge databases under-represent milder SLE cases managed in outpatient settings.³³

Healthcare resource use varied substantially across SLE manifestations in this study. LN accounted for the largest number of episodes, while less-frequent manifestations of SLE were associated with longer stays or higher costs.^43–47 Among episodes requiring ICU admission, mean ICU stays were longer than for all other hospitalisation episodes, consistent with previous reports.^48–50 These findings should be interpreted as patterns of healthcare utilisation, rather than direct indicators of disease severity or quality of care.

Admissions for renal and cardiac manifestations of SLE were more frequent in larger hospitals, likely reflecting referral pathways and healthcare organisation rather than disease distribution. Regional variation in hospitalisation rates may similarly reflect differences in diagnostic activity, referral systems and healthcare infrastructure, rather than true variation in disease burden.^{33 51 52}

The mean cost per SLE admission was lower than reported previously,⁵³ possibly because admissions with SLE recorded only as a comorbidity were excluded. Differences in costing methods, comorbidity burdens and LOSs may also explain these discrepancies.^54–57

Strengths and limitations

This nationwide study has several strengths, including broad nationwide coverage of Spanish specialised healthcare activity, standardised ICD-10 coding and analysis of clinically relevant SLE manifestations, which supported the description of nationwide patterns in episodes with SLE coded as the principal diagnosis, mortality and healthcare resource use.

However, several limitations of the study should be considered. As with all administrative databases, the MBDS may contain diagnostic misclassifications, and information on disease activity, serology and treatment was unavailable. As the analyses were conducted at the episode level, repeated admissions of the same individuals could not be identified, potentially leading to the overestimation of the number of affected individuals and influencing temporal trends. Therefore, the findings should be interpreted as episode-level measures of healthcare utilisation rather than patient-level estimates of disease burden or prognosis. Furthermore, the use of aggregated administrative data precluded adjustment for potential confounders such as age, sex, hospital characteristics and case mix. Outcomes were assessed only for healthcare episodes with SLE coded as the principal diagnosis. Although this approach improves diagnostic specificity, it may underestimate the overall burden of SLE-related hospitalisations by excluding episodes in which SLE was recorded as a secondary diagnosis, and may introduce selection bias towards admissions more directly attributable to lupus. Given the small numbers of episodes in some subgroups, the findings of the stratified analyses should be interpreted cautiously and considered exploratory due to reduced statistical precision. No adjustment for multiple comparisons was performed because these subgroup analyses were considered exploratory. Consequently, statistically significant findings in infrequent manifestations should be interpreted with caution. In addition, temporal trends may have been affected by changes in coding practices over time. Despite these limitations, this study provides nationwide estimates of hospital episodes with SLE coded as the principal diagnosis and resource use in Spain.

Conclusions

This nationwide study showed that healthcare episodes with SLE coded as the principal diagnosis in Spain increased between 2016 and 2022, whereas in-hospital mortality remained low and stable both for overall SLE and across SLE-associated manifestations. Although SLE accounted for a small proportion of overall healthcare activity, episodes with SLE coded as the principal diagnosis had longer hospital and ICU stays than all other hospitalisation episodes in unadjusted descriptive comparisons.

Resource utilisation patterns differed across SLE manifestations, with LN accounting for the largest number of episodes, while less frequent manifestations were associated with longer hospital stays or higher costs. Overall, these findings provide updated nationwide information on SLE-related hospital activity in Spain that may be relevant for healthcare planning and future epidemiological research. However, because the analyses were based on administrative episode-level data, the findings should be interpreted as measures of healthcare utilisation rather than patient-level estimates of disease burden, disease severity or prognosis.

Data availability statement

Data are available in a public, open access repository. All data relevant to the study are included in the article or uploaded as supplementary information.