Introduction

Multisystem inflammatory syndrome (MIS) is a hyper-inflammatory event that impacts multiple organ systems within the body [1]. Initially, the condition was discovered as a post-infectious complication of SARS-CoV2 in pediatric age groups exclusively and was labeled Multisystem inflammatory syndrome in children (MIS-C) [2]. Later, this syndrome was found to affect adults as well, and this variety was termed multisystem inflammatory syndrome in adults (MIS-A).

The pathophysiology of MIS involves an overactive immune system, as also seen in Kawasaki disease, while having distinct clinical characteristics that set them apart [3]. While a dysregulated immune response following infection by SARS-CoV2 plays a significant role, MIS is believed to be multifactorial, whereby both genetic and environmental factors may also impact its pathophysiology [4].

MIS-A often presents with a wide range of symptoms, including, but not limited to, fever, hypotension, dyspnea, diarrhea, signs of heart failure, and coagulation abnormalities. In their review, Patel et al. estimated that the median number of systems involved was 5 among 221 patients studied [5]. As these clinical manifestations can overlap with other conditions, and MIS-A is associated with high mortality that might also be impacted by underlying diseases and predisposing factors, a timely diagnosis and management are critical for their prognosis [1, 6, 7].

Nevertheless, there is fairly little information available regarding the demographics, risk factors, and clinical outcomes among MIS-A patients. As such, this study aims to elucidate the patient characteristics, comorbidities, and clinical outcomes associated with MIS-A to provide a better understanding of the syndrome.

Methods

We performed a retrospective study utilising data from the 2021 National Inpatient Sample (NIS) database. As per the recommendations of the CDC, we only retained patients aged ≥ 21 years [8]. The NIS is compiled each year by the Healthcare Cost and Utilization Project (HCUP), with non-federal hospitalizations, which can be extrapolated to cover 35 million national records annually. Users can study various conditions and procedures via the respective International Classification of Diseases (ICD) codes. Records before 1 October 2015 required the use of ICD-9 codes, after which the NIS transitioned to ICD-10 codes [9].

Our sample of admissions between 1 January 2021 and 31 December 2021 was stratified into groups of individuals with and without MIS-A, via the ICD-10 code “M35.81” [10]. Further patient characteristics, comorbidities, and complications were added based on recommendations from prior studies [11–14].

Baseline patient demographics and characteristics were evaluated. We used χ² tests for categorical groups, and Mann-Whitney U tests for continuous variables based on the non-parametric nature of the NIS. We further explored the presence of active COVID-19 infection, post-COVID-19 syndrome, and personal history of COVID-19 [15, 16], and multivariable regression models were used to estimate the odds of MIS-A in the presence of such risk factors. Finally, we estimated the odds of mortality associated with MIS-A and the median length of stay (LOS), as well as the hospital charge (USD).

Results

Among the total cohort of 27,948,054 patients enrolled in this investigation, 2,730 individuals were admitted with MIS-A. Several differences in characteristics about demographics and comorbidities were reported between patients with and without MIS-A (Table I). Notably, the median age of patients diagnosed with MIS-A was 58.00 years, slightly lower than the median age of 62.00 years observed in non-MIS-A admissions. In terms of age distribution, we found that 27.5% of MIS-A hospitalizations occurred in the 21–45 age group, 25.8% in the 46–60 group, and 46.7% were over 60 years old, while, amongst non-MIS-A admissions, the distribution was 28.3%, 19.5%, and 52.2%, respectively (p < 0.01). Furthermore, we also found differences in the sex, with a higher proportion of MIS-A patients being males (58.6%, p < 0.01). Additionally, there were significant differences in the racial distribution, where 56.0% of MIS-A patients were White, 20.6% Black, 16.1% Hispanic, and 3.4% Asian/Pacific Islander, compared to 65.7%, 15.5%, 12.2%, and 2.9%, respectively, in non-MIS-A admissions (p < 0.01). We also found that MIS-A hospitalizations were associated with a lower proportion of Medicare (37.1% vs. 46.8%) and a higher proportion of private medical insurance (38.2% vs. 26.9%) as opposed to those hospitalized without MIS-A (p < 0.01). MIS-A hospitalizations involved mostly urban teaching hospitals (68.9%) and hospitals in the South region (45.6%).

Patients hospitalized with MIS-A had significantly different comorbidity profiles as compared to those hospitalized without the condition, with a higher prevalence of diabetes (33.5% vs. 29.0%, p < 0.01), chronic kidney disease (CKD) (23.4% vs. 18.2%, p < 0.01), obesity (31.3% vs. 19.9%, p < 0.01) and heart failure (21.2% vs. 19.5%, p = 0.018). MIS-A had an especially strong association with autoimmune conditions (90.5% vs. 3.3%, p < 0.01). On the other hand, certain comorbid conditions were observed to report a lower prevalence in the MIS-A cohort in comparison with the non-MIS-A population. This included dyslipidemia (30.6% vs. 36.9%, p < 0.01), smoking (23.1% vs. 36.1%, p < 0.01), alcohol abuse (2.0% vs. 5.9%, p < 0.01), cancer (4.2% vs. 7.7%, p < 0.01), chronic lung disease (18.5% vs. 20.6%, p < 0.01), perivascular disease (6.1% vs. 3.8%, p < 0.01) and hypothyroidism (8.6% vs. 13.0%, p < 0.01). Hypertension was found to be marginally lower in MIS-A patients (31.3% vs. 31.4%), and this difference was observed to be statistically insignificant (p = 0.904). Our study also discovered a very strong association between active COVID-19 infection and the development of MIS-A (71.8% vs. 8.5%, aOR = 27.436, 95% CI: 25.082–30.011, p < 0.001). Interestingly, a larger number of patients with MIS-A also had a previous history of COVID-19 infection (5.3% vs. 3.3%, aOR = 2.963, 95% CI: 2.485–3.534, p < 0.01), and those with post-COVID-19-syndrome (2.0% vs. 0.1%, aOR = 32.766, 95% CI: 24.316–44.153, p < 0.001) when compared to those without MIS-A (Table II).

Regarding outcomes, MIS-A hospitalizations were significantly more likely to result in death compared to non-MIS-A hospitalizations (26.9% vs. 3.6%, aOR = 3.743, 95% CI: 3.399–4.121, p < 0.001) (Table II). Additionally, MIS-A patients tended to have longer hospital stays. The median length of hospital stays for MIS-A patients was 9.00 days (IQR: 4.00–16.00), whereas for non-MIS-A patients, it was only 3.00 days (2.00–6.00, p < 0.01). Similarly, the median hospital charge for MIS-A admissions was significantly higher than for non-MIS-A hospitalizations ($120,891, IQR: $55,528–$272,775; $40,075, IQR: $21,353–$78,747; p < 0.01).

Discussion

The primary findings from our analysis are as follows:

The findings of our study indicate a significant relationship between active, as well as prior COVID-19 infection and the development of MIS-A. This relationship held even after adjusting for confounding factors. This indicates that the suggested correlation is likely a true cause-and-effect relationship. These observations concur with an emerging literature that highlights MIS as a consequence of COVID-19 [4, 17, 18]. One major limitation of our study involves the lack of data on the type of COVID-19 virus involved. It is, therefore, vital to promote additional retrospective studies so that the genomic roles of the different variants and their roles in the pathogenesis of MIS can be better understood [19, 20].

To date, data on MIS-A have involved smaller samples, with a median age of 21 in a study by Patel et al. [5] and a mean age of 33 years by Lawrensia et al. [21], which is lower than our reported median age of 58.00. However, our sample size has allowed us to explore a wider range of patients, as compared to the 28 cases by Lawrensia et al. and 221 cases by Patel et al. [5, 21]. We also noted that our MIS-A sample was younger compared to our non-MIS-A control cohort. To evaluate the age-related immunological response and the pathophysiology of MIS-A additional studies should be encouraged.

Furthermore, the higher presence of males in our MIS-A group matches previous smaller studies of MIS-A [5]. Patel et al. associated such differences with biological risk factors as well as behavioral differences between the two sexes [5]. While past studies have associated females with a higher immune response, which predisposes them to various autoimmune conditions, the opposite differences seen among MIS-A patients warrant deeper understanding and studies [22, 23].

While a bigger sample of MIS-A patients were Whites, in comparison to non-MIS-A hospitalizations, the racial composition of patients hospitalized with MIS-A suggests a decline in the number of White patients and an increase in the numbers of Black and Hispanic patients. This study concurs with the findings of Javalkar et al., Middelburg et al., and Stierman et al. that there is a predilection for MIS among Black and Hispanic populations [24, 25]. Multiple autoimmune conditions such as SLE [26] are more common and severe in Blacks, and while a potential genetic predisposition may exist, extensive research may help bridge the gap in knowledge on the racial roles in pathogenesis and help in the treatment plans as preventive and curative measures. Finally, the higher presence of private insurance among MIS-A patients as compared to Medicare reflects the younger age in our MIS-A sample vs. our control group.

The lower presence of multiple comorbidities in our MIS-A group, such as dyslipidemia, chronic lung disease, cancer, and peripheral vascular disease, is linked with the younger age group as compared to the control cohort used. Several comorbidities seen, such as diabetes, CKD, obesity, and heart failure, have previously been strongly linked with more severe outcomes among COVID-19 patients [27–29]. In addition, 90.5% of patients with MIS-A in our study had some form of autoimmune condition. As the pathophysiology of MIS-A involves an immune dysregulation, the co-existing presence of such autoimmune conditions may put them at a much higher risk of an excessive immune response triggering MIS-A [30].

Finally, the mortality rate (26.9%) seen in MIS-A admissions in our study was higher than in past studies, such as Patel et al. who reported a mortality rate of 7% only [5]. However, our sample involved a bigger cohort, with a higher median age. Unfortunately, the NIS does not report the cause of death. Therefore, it is vital to conduct additional retrospective studies through hospital records based on various age groups and their causes of mortality.

In conclusion, we presented one of the most extensive samples of MIS-A patients and found differences in comorbidities compared to a non-MIS-A control group. Active and past COVID-19 exposures were strongly linked with a higher risk of MIS-A. Finally, MIS-A led to higher odds of all-cause in-hospital death.