Severe mental disorders (SMDs) such as bipolar and psychotic disorders affect an estimated 1–3% of the adult population [1
]. The life expectancy for individuals with SMD is reduced by approximately 10 to 20 years compared with the general population [3
]. Studies have shown that most of the reduction in life expectancy seems to be due to somatic comorbidities rather than external events such as suicide or accidents [5
]. Individuals with SMDs have an elevated risk of prematurely dying from cardiovascular disease, diabetes, and chronic obstructive pulmonary disease. The causes behind this elevated risk are not fully understood. However, some established risk factors such as smoking, substance use, obesity, poor diet, lack of exercise, and hypertension are more common in individuals with SMD [7
]. There are also concerns regarding inequality in diagnosis and treatment of somatic risk factors and enrolment in primary and secondary prophylactic measures compared with the general population [9
]. Individuals with SMDs may thus receive fewer and more delayed medical interventions. They may also be less likely to adhere to prescribed treatments and to seek medical attention for somatic diseases when needed [10
The world remains in the grip of the COVID-19 pandemic [12
]. Recent studies have shown that individuals with SMDs are at significantly increased risk of COVID-19-associated death compared with the general population [2
]. Many of the occurring somatic comorbidities in individuals with SMDs match the identified risk factors for severe COVID-19 infection. Yet, even individuals with SMD without any known risk factors seem to have a threefold increased risk of COVID-19-associated death [2
]. Some guidelines have now included SMDs as a risk group for COVID-19 [15
]. Pre-COVID-19 studies, analyzing data from 2003–2009, have indicated that individuals with SMDs have an increased risk of death associated with influenza and pneumonia [5
]. These results have not yet led to individuals with SMDs being prioritized for vaccinations against these respiratory infections [18
Therefore, we set up a population-based study to examine whether individuals with SMDs remain at a higher risk of death and hospitalization due to influenza/pneumonia. The risk of death and hospitalization associated with sepsis was also explored as it is a serious condition with similarities with COVID-19 [20
] and there are currently few studies examining sepsis in individuals with SMDs. This way, we intended to update the current evidence base as a decision aid for public health officials. Such information could motivate additional actions and strategies to promote health in these individuals, for example, by inclusion in prioritized groups for vaccination against respiratory infectious pathogens such as influenza and pneumococci. To the best of our knowledge, there are currently few other studies that have examined the risk of death associated with sepsis in general for individuals with SMDs.
This retrospective nationwide register study shows that, compared with the rest of the Swedish population, individuals with SMDs have increased risks of both death and hospitalization associated with pneumonia/influenza and sepsis. Generally, higher odds ratios were observed for death and hospitalization associated with pneumonia/influenza than for sepsis. There was an overall trend of the odds ratios peaking in the age groups 40–59 and 60–69 years, declining thereafter in the older age groups across all examined outcome categories. The smaller odds ratios in the older age groups may have arisen owing to individuals with less severe SMDs being physically healthier and having a longer life expectancy. Except for death associated with sepsis in the age groups 40–59, the highest odds ratios across all examined outcome categories were observed in individuals with SMDs without any of the known comorbidities examined in this study (diabetes, chronic lung disease, hypertension, or cardiovascular disease). However, the overall percentages in the outcome categories were generally higher in individuals with comorbidities and old age.
The increased risk of death associated with pneumonia/influenza in individuals with psychiatric disorders is in line with previous studies. Crump et al. examined causes of mortality in individuals with schizophrenia and bipolar disorders during 2003–2009. Compared with the general population in Sweden, there was an almost sevenfold increased risk of death due to influenza/pneumonia in individuals with schizophrenia and an about 3.5-fold risk in individuals with bipolar disorders [5
]. Similarly, Miller et al. in the United States (US) reported a standardized mortality ratio for pneumonia/influenza of 6.6 for patients with SMDs, defined as individuals requiring at least one inpatient psychiatric hospitalization [17
]. Standardized mortality rates for individuals with severe mental illnesses were also examined in a large retrospective cohort study in Wales [30
]. In their study, the overall standardized mortality rate for pneumonia was almost fourfold, and it was ninefold in the group 45–64 years. For sepsis, the standardized mortality rate was threefold. To the best of our knowledge, there are currently few other studies that have examined the risk of death associated with sepsis in general for individuals with SMDs.
The risk of postoperative sepsis and associated death was examined in the United States. In that study, patients with schizophrenia had double the odds of postoperative sepsis compared with patients without schizophrenia. The risk of death was increased sevenfold [31
]. A Danish study examined the risk of death within 30 days after infection. In that study, patients with psychotic or bipolar disorders had around 30 percent increased risks of death due to either pneumonia or sepsis compared with the general population [32
]. Adverse clinical outcomes among patients hospitalized for pneumonia with and without schizophrenia were examined in a study from Taiwan. In that study, patients with schizophrenia had a 1.3- to 1.8-fold increased risk of ICU admission, mechanical ventilation, and acute respiratory failure [33
]. However, the risk of in-hospital death did not differ significantly.
This study has several strengths and limitations. By including the entire Swedish population in the analysis, the risk of selection bias was eliminated, and statistical power was brought to a maximum. The results are thus valid for the Swedish population, but additional studies are needed to evaluate the generalizability to other parts of the world. All data were summarized by a statistician at the Swedish National Board of Health and Welfare, independently from the research group, thereby eliminating the risk of observation bias. The registers used in this study are regarded as highly validated [23
]. The combination of influenza and all-cause pneumonia into a single category enabled comparison with other studies of individuals with SMDs such as the studies by Crump et al. [5
], but also constitutes a limitation as differentiation between specific infectious agents is not possible. As sepsis was identified by ICD-10 codes and not laboratory records, reliable data regarding etiology were not available for the current data set. The most common bacterial pathogen for pneumonia is Streptococcus pneumoniae
and one study has reported that individuals with schizophrenia or bipolar disorder are at increased risk of both pneumococcal pneumonia and septicemia [34
]. Information regarding the causative infectious agent would, for example, be valuable for advising whether vaccination against influenza and/or pneumococci is motivated. Guidelines for influenza vaccination recommend prioritization for everyone above 65 years of age in Europe and above 50 years of age in the United States regardless of other risk factors. Therefore, stratification of age taking account of these age thresholds could have provided more specific information on those currently not prioritized for influenza vaccination [18
]. Furthermore, the combination of psychotic and bipolar disorder into a single SMD variable at the point of ordering the data prevented separate analyses of the disorders. Nevertheless, an increased risk of death associated with influenza/pneumonia in individuals with SMDs remains in this analogous follow-up to Crump et al. In future research, SMDs should be explored further, by stratification by medication adherence, psychiatric admissions, and use of mental health legislation. Illness duration is another potentially important factor. However, such a more detailed study of SMDs was not possible with the data set available for the current study. We chose to not include depressive disorders into the group of SMD owing to difficulties to quantify severity in this heterogeneous group. Many of the individuals with major depressive disorders are monitored by primary care, and thus not with full certainty included in the Swedish National Patient Register.
There could be also other contributing factors to the results, the exploration of which is beyond the scope of the current study. Important risk factors such as obesity, socioeconomic status, smoking, and excessive alcohol use are not recorded reliably, if at all, in the registers [7
]. Owing to the limitations of the registers and the retrospective nature of the current study, further stratification by characteristics or matching of study populations was not possible. Prospective studies are needed in future to reduce confounding. The list of comorbidities to be explored could be expanded in future research. Dementia is one such example, recently reported to be much more prevalent among patients with schizophrenia. Dementia is also a known risk factor for pneumonia [46
]. There are also other socioeconomic and environmental differences; individuals with SMDs are more likely to be homeless, unemployed, and live in poverty, all of which may affect access to healthcare and increase the risk of infection [48
]. Individuals with SMDs may also receive fewer and more delayed medical interventions for somatic diseases, possibly associated with discrimination and stigmatization [10
]. Other undiagnosed somatic comorbidities may be more prevalent in individuals with SMDs, and may thus contribute to adverse outcomes [9
]. For any given comorbidity, individuals with SMDs may also have greater clinical severity, for example, uncontrolled diabetes [57
]. These unexplored factors most likely also add to the differences in the outcomes of this study, although the magnitudes are difficult to estimate. Furthermore, psychiatric medications commonly have weight gain as a side-effect and some may affect immune function, also likely affecting the outcomes [58
Finally, the current COVID-19 pandemic has sparked extensive research and discussion regarding which individuals should be prioritized for vaccination [61
]. Emerging data indicate that individuals with SMDs are at increased risk for severe COVID-19 infection and several guidelines have recently included SMDs as a high-risk group for COVID-19 [15
]. To the best of our knowledge, SMD without other known chronic medical diseases is not considered as a potential risk group for other respiratory infections such as severe influenza [18
]. The European Centre for Disease Prevention and Control defines risk groups as persons at higher risk of adverse outcomes when infected with influenza and for whom vaccines are demonstrated to reduce the risk of those outcomes [63
]. Additional studies are needed to confirm the influenza virus as a definitive cause of the increased risks of death and hospitalization and to explore the potential effects of vaccines against influenza in individuals with SMD. As with COVID-19, the findings of this study suggest that coexisting somatic comorbidities not are enough to explain the increased risks for individuals with SMDs associated with influenza/pneumonia. Therefore, the argument that individuals with SMDs will already be covered by vaccination priority strategies because of their physical health status does not hold.