Rising incidence of systemic autoimmune inflammatory rheumatic diseases during the COVID-19 pandemic: a geographical cohort study

Abstract

Objective

The natural infection with SARS-CoV-2, or vaccination against it, has been postulated to directly contribute to an increase in the incidence of autoimmune inflammatory rheumatic diseases (AIIRDs). Conversely, preventive measures limiting access to healthcare services could have resulted in missed or delayed AIIRD diagnoses or have reduced the infection rate of any triggering infections. We aimed to define real-life trends in AIIRD diagnoses from the prepandemic period through 2023 in a large and geographically circumscribed population of 6.5 million inhabitants.

Methods

AIIRDs’ annual diagnosis rates from 2017 to 2023 were derived from the registration of disease-specific exemption codes of the resident population of Lazio, a highly populated region in central Italy. Incidence rate ratios (IRRs) were calculated to compare pandemic and average prepandemic rates (2017–2019). Poisson regression was used to define statistically significant changes.

Results

A total of 16 254 AIIRD diagnoses were registered over the 7-year period. The average prepandemic incidence of AIIRDs was 4.81 per 10 000 inhabitants (95% CI 4.69 to 4.92). Compared with the prepandemic period, the diagnosis rate decreased in 2020 (IRR 0.68, 95% CI 0.64 to 0.72) but remained above prepandemic levels in 2021, 2022 and 2023. In 2023, there was a 22% increase in AIIRD incidence compared with prepandemic levels (IRR 1.22, 95% CI 1.17 to 1.28, p<0.001). This excess incidence was primarily driven by increases in both primary arthritides and systemic rheumatic diseases.

Conclusions

We observed a temporary decline in diagnosis in 2020, followed by a substantial increase from 2021 to 2023. This trend (decline and increase) may be linked to COVID-19 infection or to the reduction and subsequently potential increase of other infectious triggers following the use of preventive measures, such as facial masks and social distancing.

WHAT IS ALREADY KNOWN ON THIS TOPIC

• It is well known that infections can trigger autoimmune inflammatory rheumatic diseases (AIIRDs) in genetically predisposed individuals, and a similar mechanism has been debated for antigen exposure related to vaccination. Most of the data supporting a potential link between SARS-CoV-2 antigen exposure and the onset of AIIRDs comes from individual case reports and smaller case series. It remains unclear whether the pandemic led to an increase in new AIIRD cases above the expected rate or whether other triggers might have driven the onset of these conditions.

WHAT THIS STUDY ADDS

• Using administrative data, we calculated the incidence of nine AIIRDs from 2019 to 2023 in a region of central Italy with 6.5 million inhabitants. We observed a significant increase across primary arthritides and systemic rheumatic diseases. The increase ranged from 22% to 65%, depending on the specific disease.

HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

• A substantial increase in the number of new AIIRD cases may be anticipated in the coming years in Western countries. These findings should inform clinicians and guide public health decision-making and planning efforts.

The SARS-CoV-2 was identified at the end of 2019 as the aetiological agent of COVID-19, a new form of infection that rapidly and unexpectedly spread worldwide. Italy has been among the most severely involved countries. The WHO declared COVID-19 a pandemic at the beginning of 2020 and declared it over in May 2023.

Both natural SARS-CoV-2 infection and COVID-19 vaccination have also been scrutinised for their potential role in triggering autoimmune inflammatory rheumatic diseases (AIIRDs) in predisposed patients. The notion that AIIRDs can be triggered by SARS-CoV-2 antigen exposure is mainly supported by case reports and series, while larger epidemiological studies are limited and substantially conflicting.¹

This contradictory evidence may stem from several factors. First, public health strategies adopted to curb the virus’s spread may have significantly confounded results, as delays in diagnosing new AIIRDs could have masked the true impact of COVID-19.² Second, nearly the entire at-risk population in Western countries has been exposed to SARS-CoV-2 antigens through natural infection and/or vaccination, with many individuals experiencing multiple infections or vaccinations. Third, the latency period between a triggering SARS-CoV-2 infection and the onset of AIIRDs remains undefined, making the assessment of temporality criteria to support a causal relationship arduous. Finally, substantial disparities emerged in the course of COVID-19 with ethnic minority groups, urban areas and low-income populations experiencing disproportionately higher infection rates and worse prognosis.

Importantly, reporting a statistical association between COVID-19 and the onset of AIIRDs cannot determine whether predisposed patients would have developed these diseases regardless of infection or vaccination, or if it represents a real diagnostic excess over the expected rate of new diagnoses. The aim of our study was, therefore, to investigate a diagnostic excess rate of AIIRDs during the pandemic period (from 2020 to 2023) and compare it to the incidence of the same diseases in the years prior to the COVID-19 outbreak (from 2017 to 2019), in Lazio, a large and highly populated administrative region of central Italy.

Patients and methods

We performed a retrospective geographical cohort analysis of the Italian Lazio region from 2017 to 2023.

Lazio is 1 of the 20 administrative regions of Italy, situated in the central peninsular section of the country. This region has about 6.5 million inhabitants, making it the second most populated region in Italy, and includes the metropolitan area of Rome, an urban area with one of the highest population densities in the country. With more than 2 million registered COVID-19 cases, it was one of the most severely affected regions during the pandemic.

The subjects included in the study were uniquely identified by their Italian tax code. The diagnosis rate was derived from the registration of new patient disease-specific exception codes. In the Italian healthcare system, disease-related exception codes refer to specific alphanumeric codes assigned to individuals who are exempt from paying healthcare service costs due to a specific clinical condition. These exemptions can apply to various services, including specialist visits, diagnostic tests and treatments.³ The AIIRDs diagnoses considered in our analysis were rheumatoid arthritis (RA), psoriatic arthritis (PsA), axial spondylarthritis (axSpA), systemic lupus erythematosus (SLE), primary Sjögren’s syndrome (SjS), systemic sclerosis (SSc), idiopathic inflammatory myositis (IIM), undifferentiated connective tissue disease (UCTD) and systemic vasculitis (SV). The IIM group included diagnoses of polymyositis, dermatomyositis, inclusion body myositis, eosinophilic myositis, mixed connective tissue disease and anti-synthetase syndrome. The SV group included diagnoses of giant cell arteritis, takayasu arteritis, polyarteritis nodosa, microscopic polyangiitis, granulomatosis with polyangiitis, eosinophilic granulomatosis with polyangiitis, Behçet’s disease, cryoglobulinaemic vasculitis and IgA vasculitis (former Henoch-Schönlein purpura).

Annual incidence rates (IRs) with 95% CIs were calculated from new AIIRD diagnosis, and the at-risk population was evaluated at the beginning of each year. IR ratios (IRRs) with 95% CIs were calculated comparing each pandemic year to the average of the 2017–2019 period. Statistical significances of changes in follow-up were determined using Poisson regression modelling. P values less than 0.05 were regarded as statistically significant. Statistical analyses were carried out using R Studio (V.2023.12.1).

Results

A total of 16 254 diagnoses of AIIRDs were registered over 7 years in a population of about 6.5 million inhabitants; 9468 of these diagnoses were performed from 2020 to 2023. The total diagnoses include 5172 cases of RA, 2950 of PsA, 1296 of SpA, 2129 of UCTD, 11577 of SjS, 884 of SLE, 1006 of SSc, 378 of IIM and 862 of SV. Annual total and disease-specific IRs from 2017 to 2023 are plotted in figures1 2 and detailed in table 1, along with the corresponding IRRs comparing pandemic years to 2019.

Figure 1. Overall AIIRD annual incidence from 2019 to 2022. Red dashed line indicates average 2017–2019 rate. **Rheumatology and Clinical Immunology, ***p<0.001 compared with 2017–2019. AIIRD, autoimmune inflammatory rheumatic disease.

Figure 2. Single AIIRD annual incidence from 2019 to 2023. Red dashed line indicates average 2017–2019 rate. *p<0.05, **p<0.01, ***p<0.001 compared with 2017–2019. AIIRD, autoimmune inflammatory rheumatic disease; IIM, idiopathic inflammatory myositis; PsA, psoriatic arthritis; RA, rheumatoid arthritis; SjS, Sjogren’s syndrome; SLE, systemic lupus erythematosus; SpA, spondyloarthritis; SSc, systemic sclerosis; SV, systemic vasculitis; UCTD, undifferentiated connective tissue disease.

Table 1. IRs and IRRs of AIIRDs from 2017 to 2023.

Disease	2017 IR per 10 000(95% CI)	2018 IR per 10 000(95% CI)	2019 IR per 10 000(95% CI)	2017–2019 IR for 10 000(95% CI)	2020 IR per 10 000(95% CI)	IR 2021per 10 000(95% CI)	IR 2022per 10 000(95% CI)	IR 2023per 10 000(95% CI)	IRR 2020vs 2017–2019(95% CI)	IRR 2021vs 2017–2019(95% CI)	IRR 2022vs 2017–2019(95% CI)	IRR 2023vs 2017–2019(95% CI)
AIRRS overall	4.44(4.25 to 4.63)	5.05(4.85 to 5.26)	4.93(4.73 to 5.14)	4.81(4.69 to 4.92)	3.27(3.10 to 3.43)	5.97(5.75 to 6.20)	5.20(4.99 to 5.41)	5.88(5.66 to 6.10)	0.68(0.64 to 0.72)	1.24(1.19 to 1.30)	1.08(1.03 to 1.13)	1.22(1.17 to 1.28)
RA	1.46(1.35 to 1.57)	1.47(1.36 to 1.58)	1.50(1.39 to 1.62)	1.48(1.42 to 1.54)	1.09(1.00 to 1.19)	1.93(1.81 to 2.06)	1.77(1.66 to 1.90)	1.82(1.70 to 1.95)	0.74(0.70 to 0.81)	1.30(1.21 to 1.41)	1.20(1.11 to 1.30)	1.23(1.14 to 1.34)
SpA	0.30(0.25 to 0.35)	0.36(0.31 to 0.42)	0.39(0.33 to 0.44)	0.35(0.32 to 0.38)	0.25(0.21 to 0.30)	0.54(0.49 to 0.61)	0.43(0.38 to 0.50)	0.49(0.43 to 0.56)	0.71(0.58 to 0.87)	1.54(1.32 to 1.69)	1.23(1.04 to 1.45)	1.40(1.20 to 1.64)
PsA	0.82(0.74 to 0.90)	0.92(0.83 to 1.01)	0.86(0.78 to 0.95)	0.87(0.82 to 0.91)	0.61(0.55 to 0.69)	1.14(1.04 to 1.24)	0.88(0.80 to 0.97)	1.08(0.98 to 1.17)	0.71(0.62 to 0.81)	1.31(1.18 to 1.45)	1.01(0.91 to 1.13)	1.24(1.12 to 1.38)
SLE	0.28(0.23 to 0.32)	0.26(0.22 to 0.31)	0.28(0.24 to 0.33)	0.27(0.25 to 0.30)	0.21(0.17 to 0.25)	0.35(0.30 to 0.41)	0.24(0.20 to 0.29)	0.26(0.21 to 0.31)	0.75(0.60 to 0.94)	1.28(1.07 to 1.54)	0.89(0.72 to 1.10)	0.94(0.77 to 1.16)
SjS	0.41(0.36 to 0.47)	0.44(0.38 to 0.42)	0.47(0.41 to 0.53)	0.44(0.41 to 0.47)	0.31(0.26 to 0.36)	0.52(0.46 to 0.60)	0.58(0.52 to 0.66)	0.63(0.56 to 0.71)	0.70(0.58 to 0.84)	1.20(1.03 to 1.39)	1.33(1.15 to 1.53)	1.44(1.25 to 1.65)
SSc	0.23(0.18 to 0.27)	0.51(0.44 to 0.57)	0.34(0.29 to 0.40)	0.36(0.33 to 0.39)	0.31(0.26 to 0.36)	0.53(0.46 to 0.60)	0.58(0.52 to 0.66)	0.35(0.29 to 0.40)	0.49(0.39 to 0.62)	0.73(0.60 to 0.89)	0.79(0.66 to 0.96)	0.97(0.81 to 1.16)
IIM	0.07(0.05 to 0.09)	0.11(0.08 to 0.14)	0.13(0.10 to 0.16)	0.10(0.09 to 0.12)	0.07 (0.05 to 0.10)	0.10(0.08 to 0.14)	0.16(0.13 to 0.20)	0.17(0.13 to 0.21)	0.71(0.49 to 1.04)	1.01(0.73 to 1.40)	1.56(1.18 to 2.06)	1.65(1.25 to 2.17)
SV	0.34(0.29 to 0.39)	0.27(0.23 to 0.32)	0.26(0.21 to 0.30)	0.29(0.26 to 0.32)	0.17 (0.14 to 0.21)	0.26 (0.22 to 0.31)	0.26(0.22 to 0.31)	0.28(0.23 to 0.33)	0.58(0.46 to 0.74)	0.89(0.83 to 1.09)	0.90(0.74 to 1.10)	0.96(0.79 to 1.17)
UCTD	0.53(0.46 to 0.59)	0.71(0.63 to 0.78)	0.70(0.62 to 0.77)	0.64(0.60 to 0.69)	0.38(0.33 to 0.44)	0.86(0.78 to 0.95)	0.58(0.51 to 0.65)	0.79(0.71 to 0.88)	0.59(0.50 to 0.69)	1.34(1.19 to 1.50)	0.90(0.78 to 1.03)	1.23(1.09 to 1.39)

AIIRD, autoimmune inflammatory rheumatic disease; IIM, idiopathic inflammatory myositis; IR, incidence rate; IRR, IR ratio; PsA, psoriatic arthritis; RA, rheumatoid arthritis; SjS, primary Sjögren's Syndrome; SLE, systemic erythematosus lupus; SpA, spondyloarthritis; SSc, systemic sclerosis; SV, systemic vasculitis; UCTD, undifferentiated connective tissue disease.

The diagnostic rates for overall AIIRDs were first assessed. In the prepandemic period (2017–2019), the average incidence of AIIRDs was 4.81 per 10 000 inhabitants (95% CI 4.69 to 4.92). Compared with the prepandemic period, the overall diagnostic rate significantly decreased in 2020 (IRR 0.68, 95% CI 0.64 to 0.72, p<0.001). However, rates remained above prepandemic levels in 2021, 2022 and 2023. By the final year of follow-up, the overall incidence had risen to 5.88 per 10 000 inhabitants (95% CI 5.66 to 6.10), representing a 22% statistically significant increase compared with the prepandemic levels (IRR 1.22, 95% CI 1.17 to 1.28, p<0.001).

As expected, RA had the highest incidence among AIIRDs at all time points, and its diagnostic trend appears to drive the overall incidence rise observed for AIIRDs as illustrated in figure 2. Compared with the prepandemic incidence (1.48 per 10 000 inhabitants, 95% CI 1.42 to 1.54), the incidence of RA declined in 2020 but remained consistently above baseline levels in 2021, 2022 and 2023. By the final year of follow-up, the incidence had increased by 23% (IRR 1.23, 95% CI 1.14 to 1.34, p<0.001). A similar pattern was also observed for axSpA and PsA: after a temporary decline in 2020, both conditions saw a rebound peak in 2021 and still showed statistically significant increases in 2023 compared with the prepandemic levels. The incidence rose by 40% for AxSpA (IRR 1.40, 95% CI 1.20 to 1.564, p<0.001) and 24% for PsA (IRR 1.24, 95% CI 1.20 to 1.64, p<0.001).

Among systemic rheumatic diseases, SjS displayed a distinct upward trend after 2020, reaching a 44% increase (IRR 1.44, 95% CI 1.25 to 1.65, p<0.001) in 2023 compared with 2017–2019, following a temporary reduction observed in 2020. Similarly, a net increase in disease incidence in 2023 compared with the prepandemic levels was also observed for UCTD (IRR 1.23, 95% CI 1.09 to 1.39, p<0.01) and IIM (IRR 1.65, 95% CI 1.25 to 2.17, p<0.01). In contrast, SSc and SLE did not show statistically significant increases in 2022 or 2023 compared with the prepandemic period.

Discussion

This observational study evaluated changes in the diagnostic rates of AIIRDs over the 4 years of SARS-CoV-2 outbreak, compared with the prepandemic baseline. We found that the overall AIIRD diagnosis rate declined in 2020, followed by an increase in the subsequent pandemic years. This increase was more pronounced in 2021 and largely persisted in 2022 and 2023 for both primary arthritides and systemic rheumatic diseases, exceeding the expected rebound due to possible missed diagnoses during the first year of the pandemic (2020). Among the studied conditions, only SLE and SSc returned to the prepandemic levels by 2023.

Our data can be considered very robust since the study cohort was geographically determined with all information derived from administrative sources, ensuring that diagnoses made in both rheumatology hubs and spokes were included. This approach helped to prevent possible selection biases, providing comprehensive data coverage of the diagnostic rates within a large region.

Unlike cross-sectional, case–control or cohort studies, this time series analysis of new AIIRD diagnosis rates within a geographical cohort cannot infer a causal relationship with COVID-19. However, it can identify real-life changes in diagnostic rates during the pandemic, potentially reflecting the combined effects of infection, vaccination and non-pharmaceutical COVID-19 measures. By using the 2017–2019 period as the comparator, we likely excluded cases that would have emerged in predisposed individuals due to triggers other than SARS-CoV-2.

Time-series data on AIIRD incidence were previously published in South Korea and in the UK. Ahn and colleagues compared AIIRD incidence in 2020 with the same incidence in the previous 3 years, using data derived from the South Korean Health Insurance Review and Assessment Service database.⁴ Russell et al reported AIIRD incidence from the second half of 2019 to March 2022 in a cohort of 17 million British patients, using the OpenSAFELY data platform across the country.⁵ Some conflicting findings between those two cohorts and our data can be acknowledged.

The reduced diagnostic rate across AIIRDs observed in our study, during the first year of the pandemic, is similar to the decrease observed in RA, axSpA and PsA incidences in the British cohort, as well as to the decline of the incidence of IIM, SLE and Behçet’s disease reported in the South Korea analysis. However, the Korean group did not report a significant reduction of the diagnostic rates of RA, axSpA, SjS and SSc compared with the prepandemic levels. Moreover, the follow-up was extended to 2021 only in the British cohort, where, unlike our observations, a rebound in AIIRD diagnostic rates was not observed.

Limitations in patient access to rheumatology services due to public health strategies aimed at limiting the virus spread may have artificially reduced the rate of new diagnoses in 2020. These diagnoses could have progressively increased in 2021 with the reorganisation of healthcare services. This hypothesis is supported by the demonstrated transient reduction in rheumatology healthcare service utilisation in 2020,⁶ and by similar trends reported in non-rheumatic diseases.⁷ Country-specific factors, such as the reorganisation of healthcare services and diagnosis registration modality, could explain the heterogeneous observations across different British, Korean and Italian contexts. The British group, in particular, cited shorter in-person assessment times and an increased use of virtual consultations in 2020, which may have blunted the expected 2021 diagnostic rebound.

An alternative or complementary hypothesis is that the use of facial masks and social distancing could have prevented viral infections different from COVID-19 that may have been implicated in the development of autoimmune diseases.⁸ This preventive effect was progressively lost with the loosening of these prophylactic strategies after 2021 and could have contributed to the diagnostic rebound. Notably, although the Italian government adopted the first restrictive measures 4 weeks earlier than the British did, the lockdown lasted for the same duration in both countries, while other containment measures (eg, face mask use) were relaxed later in Italy.⁹ The absence of a 2021 increase in RA, axSpA and PsA diagnoses in the UK after the 2020 drop cannot, therefore, be attributed to stricter British restrictions, but probably to still reduce access to rheumatology consultation.

COVID-19 vaccination has been blamed for a short-term increased risk of AIIRDs, particularly of vasculitis,¹⁰ in many case reports or series. Our data do not support this hypothesis since SV did not peak with the beginning of the vaccination campaign. These data align with the safety reports of the vaccination in AIIRD patients¹¹ and with the lack of association or even with the protective role of COVID-19 vaccination towards the development of new AIIRDs reported in large Asian cohorts.^{12 13} Conversely, the increase in IIM incidence could align with the concerns about the typology of COVID-19 infection, or the vaccine exposure, and IIM.¹⁴ Further information is needed to support this observation and demonstrate a causative association, with a more comprehensive risk–benefit analysis.

Our study extended the follow-up to 4 years, reporting a net increase in RA, PsA, SpA, UCTD, SV diagnoses in 2023 compared with 2017–2019. This indicates an excess over the expected incidence. Comparable follow-up time series data are not available, but this is consistent with longitudinal controlled studies reporting an increase in RA incidence after SARS-CoV-2 infection.¹⁵ Data about PsA, SpA, SjS, UCTD are scarce, with our study reporting for the first time a statistically significant increase in the diagnostic rate during the pandemic.

Some limitations must be considered in the generalisation of the information derived from administrative data since the diagnosis rate may not coincide with biological disease incidence. As discussed, diagnosis registration modalities could differ according to national or local legislations and be affected by patient access to healthcare services. Moreover, some patients might decide not to register their exception codes because they could benefit from other exemptions to healthcare costs due to concomitant diagnoses, unemployment or low salary. Notably, the exemption codes for SSc and UCTD were updated in 2017, which could have resulted in an artificial diagnostic excess during the first year of the prepandemic observation period, potentially affecting the results.

In conclusion, we observed real-life disease-specific trends in AIIRDs diagnosis during the pandemic as a potential result of COVID-19 infection, preventive measures and vaccination in a large geographically circumscribed cohort of several diseases simultaneously. The increase in diagnostic rates was persistent for RA, SpA, PsA, SjS, UCTD, IIM and SV. SLE and SSc did not show a statistically more frequent incidence during the pandemic.

These data could inform public health decision-making and social planning. Follow-up needs to be extended to evaluate any long-term effects on AIIRD diagnostic rates that could be related to the COVID-19 pandemic.