Смертность и причины смерти для лиц с множественным склерозом: исследование, проведенное финским национальным регистром.

Introduction

Multiple sclerosis (MS) is an inflammatory disease of the central nervous system (CNS) characterized by inflammation, demyelination, and axonal loss. Although significant advances in MS treatment have been accomplished, it remains one of the most common causes of neurological disability in young adults globally. MS is associated with functional loss [ 1 ], early unemployment [ 2 , 3 ], and early death; on average, patients with MS (pwMS) have shortened life expectancies by 7–14 years [ 4 – 10 ]. The specific reasons for this premature death remain unclear.

During recent decades, MS epidemiology and treatment have changed in ways that could affect mortality. Incidence has increased [ 11 , 12 ] and numerous new disease-modifying therapies (DMTs) emerged. With increased use of MRI, developing diagnostic criteria, and improved access to neurological health service, the additional detection of benign cases could result in increasing survival, as well as improved care, rehabilitation, and treatment of symptoms and comorbidities [ 5 , 12 , 13 ]. At the same time, the lifespan of the general population has increased. The changes in MS mortality over time, and how these changes compare to trends in mortality in the general population are incompletely understood.

Standardized mortality ratio (SMR) serves as a measure of relative mortality risk. A meta-analysis by Smyrke et al., covering a period from 1949 to 2013, found no support of a reduced all-cause SMR for pwMS compared to counter-parts over the last 65 years [ 14 ]. Accordingly, mortality has decreased among pwMS at similar rates to the general population, possibly due to advances in modern medicine and lifestyle improvements. Although this is consistent with a previous systematic review by Manouchrinia et al., including 12 studies covering the period 1949–2012 [ 15 ], it contrasts with several single-population-based studies from the Nordic countries. A rise in survival in MS over time has been observed in Norway [ 5 , 6 ], in Denmark, [ 12 ] and in Sweden [ 16 ]. A study on pwMS in New Zealand also found that MS survival had increased relative to the general population by over 15 years [ 17 ].

A previous study from Finland showed threefold higher mortality in pwMS diagnosed 1964–1993 in comparison to controls [ 10 ]. MS was the underlying cause of death in most of the patients, while infections, gastrointestinal causes, and suicides also contributed to excess mortality. The purpose of this study was to perform an updated investigation at a national level data on all-cause and cause-specific mortality among Finnish pwMS, as well as the changes in mortality rates over time. Comparison was made to national aggregated population data from Statistics Finland.

Methods

Study cohort

For this nationwide register-based cohort study, pwMS alive since 1 st January 1971 until the end of 2019 were identified from the Finnish MS-register and the nationwide hospital discharge register HILMO, which started data collection in 1969 (Fig.  1 ). HILMO includes data from Hospital Discharge Register during 1969–1993, Care Register for Health Care during 1994–2020, and Register for Primary Health Care Visits during 2011–2021. The Finnish MS registry was launched in January 2014. It is integrated with all the largest hospitals’ electronic patient record systems and with both prospective and retrospective data collection covers now data of up to 90% of Finnish pwMS [ 18 ]. Deaths and causes of death were collected since 1 st January 1971 until 31 st December 2020 from Statistics Finland and Population Register Centre's Population Information System. Findata approval (permit number THL/623/14.02.00/2021) for the study was obtained, and the data were stored and analyzed within the data secure environment of the Wellbeing County of Southwest Finland, Atolli.

Forming the study population. PwMS were identified and included from both MS-register (prioritized) and national care register (complementing search) starting from the beginning of the year 1971. The following exclusion criteria were used: only one documented visit ( n  = 636 for the national care register data), death not observed and last documented record before the year 1971 ( n  = 15 for MS-register data, n  = 0 for care register data), and diagnosis year and year of death the same ( n  = 11 for MS-register data, n  = 0 for care register data). The study population was thus 16 602 pwMS

Ethical approval

According to Finnish law, ethics committee approval was not required, since the study was based on administrative register data and did not involve any contact with patients. The study was approved by the Turku and Helsinki University Hospital Research Services. The data processing practices followed the EU Data Protection Directive rules.

Standardized mortality ratio and excess death rate

Yearly Standardized mortality ratios (SMR) were determined by dividing the number of observed deaths among pwMS by the number of expected deaths in the general population matched by sex, age group, and calendar-year. For longer time periods, Poisson regression was used for SMR calculation. Excess death rate (EDR) was determined by the difference between the observed number of deaths per 1000 person-years and the corresponding expected number. For the data on general population, we used national aggregated population data from Statistics Finland. To avoid bias caused by sharp MS incidence and prevalence changes, we analyzed SMR for 20-year periods of follow-up and used additionally EDR that is considered by some authors a better parameter for longitudinal mortality analyses [ 6 ].

Causes of death were grouped similar to what was used in [ 19 ] and cause-specific SMRs (CS-SMR’s) are presented for the latter observation period of 2000–2020. CS-SMRs were calculated using Opensource Statistics Finland yearly death count data for sex and each ICD-10 code.

Charlson comorbidity index (CCI)

To assess for the effect of comorbidities on mortality in pwMS, we calculated the CCI as described previously [ 20 ]. To analyze the effect of comorbidities on mortality, we compared the CCI for pwMS deceased during 2000–2020 to those alive at the end of 2020.

Life expectancy and survival from diagnosis

Median life expectancies were calculated for both observed data and compared to the general population. Life tables provided by Human Mortality Database were utilized to calculate expected life expectancies. Also, survival from date of MS diagnosis was assessed fitting survival curves since corresponding diagnosis date to death or end of follow-up.

Statistical analyses

Partial dates for MS-register data were imputed as middle of the year or month, where applicable. Descriptive analyses were conducted using summary statistics. Numerical variables were expressed as means with standard deviations (SDs) or medians with interquartile ranges (IQRs). Categorical variables were expressed as frequencies and proportions based on non-missing data.

15-year survival for sex, diagnosis age, and diagnosis year was analyzed fitting Proportional Hazards Regression model using Efron approximation for ties. If death was not observed, patients were censored at 15-year mark or at final day of 2020, which occurred first. Median observed life expectancies were calculated fitting survival curves since birth until death or final day of 2020. Median expected life expectancies were calculated utilizing sex and birth year matched life tables where expected survival was estimated using Ederer method. Survival from diagnosis was analyzed based on survival curves since date of diagnosis to death or final day of 2020. Survival analyses were limited for diagnosis years between 1975 and 2019.

Standardized mortality ratios were calculated using yearly death rate data for 5-year age groups, sex, and calendar-year. SMRs for longer time periods were estimated using Poisson regression, where number of observed deaths was modeled with logarithm-transformed expected deaths as offset. EDR per 1000 person-years was calculated utilizing difference between observed and expected death, and corresponding population at risk. Confidence intervals for EDRs were calculated using normal approximation.

P values under 0.05 were considered as significant. RStudio (Version 2023.06.01) was used for all data modification and analysis.

Results

In our study population of 16 602 pwMS, median age at MS onset was 36.0 years (Q1–Q3 28.0–46.0) and majority (69.7%) had been diagnosed after 1995 (Table  1 ). For this study population, a total of 4124 deaths were recorded starting from the year 1971. The range of the follow-up was 1 to 45 years. Forming the study population is demonstrated in Fig.  1 .

Table 1 Demographics of the study cohort, number of deaths recorded, age at death, and Charlson Comorbidity Index for all pwMS identified in the study Total population n  = 16,602 Females n  = 11,324 (68.2%) Males n  = 5278 (31.8%) Age at MS onset*; median (Q1-Q3) 36.0 (28.0–46.0) 35.0 (27.0–45.0) 37.0 (28.0–47.0) Age at MS diagnosis; median (Q1-Q3) 39.0 (30.0–48.0) 38.0 (30.0–48.0) 39.0 (31.0–49.0) Diagnosis year; n
(%) Before 1975 1062 (6.4%) 626 (5.5%) 436 (8.3%) 1975–1985 1842 (11.1%) 1137 (10.0%) 705 (13.4%) 1986–1995 2122 (12.8%) 1433 (12.7%) 689 (13.1%) 1996–2005 4716 (28.4%) 3319 (29.3%) 1397 (26.5%) 2006–2019 6860 (41.3%) 4809 (42.5%) 2051 (38.9%) Number of deaths Before 1980 188 80 108 1980–1989 511 260 251 1990–1999 739 429 310 2000–2009 1048 615 433 2010–2020 1638 1030 608 Observed Charlson Comorbidity Index (CCI); median (Q1–Q3) Died between 2000 and 2020 1.00 (0.00–3.00) 1.00 (0.00–3.00) 1.00 (0.00–3.00) Alive at end of 2020 0.00 (0.00–1.00) 0.00 (0.00–1.00) 0.00 (0.00–1.00) * Missing MS onset dates imputed based on 10-year diagnosis interval groups

Demographics of the study cohort, number of deaths recorded, age at death, and Charlson Comorbidity Index for all pwMS identified in the study

* Missing MS onset dates imputed based on 10-year diagnosis interval groups

Overall mortality between 1980 and 2020 was higher for pwMS compared to the general population, with an SMR of 2.40 (95% CI 2.32, 2.47; p  < 0.001) for all patients and 2.65 (95% CI 2.55, 2.76; p  < 0.001) for females (Table  2 ). During the whole study period of 1980–2020, EDR was 9.03 (95% CI 8.74, 9.32). During 1980–1999, SMR for pwMS was 3.07 (95% CI 2.91, 3.25; p  < 0.001) and EDR 14.05 (95% CI 13.72–14.37), and during 2000–2020 2.18 (95% CI 2.10–2.26) and 7.48 (95% CI 7.2–7.75), respectively. SMR was higher for female pwMS and EDR for male patients at both periods (Table  2 ).

Table 2 Mortality for pwMS calculated by SMR and EDR based on overall mortality during 1980–2020 for females and males, and 15-year mortality from diagnosis by both diagnosis year and age at diagnosis Observed Expected SMR (95% CI) EDR (95% CI) Overall mortality 1980–2020  All patients 3936 1640 ***2.40 (2.32, 2.47) 9.03 (8.74, 9.32)  Females 2334 880 ***2.65 (2.55, 2.76) 8.26 (7.99, 8.52)  Males 1602 760 ***2.10 (2.00, 2.21) 10.76 (10.42, 11.10) Overall mortality 1980–1999  All patients 1250 407 ***3.07 (2.90, 3.24) 14.05 (13.72, 14.37)  Females 689 194 ***3.55 (3.30, 3.83) 12.67 (12.37, 12.96)  Males 561 213 ***2.62 (2.42, 2.85) 16.63 (16.25, 17.01) Overall mortality 2000–2020  All patients 2686 1233 ***2.18 (2.10, 2.26) 7.48 (7.20, 7.75)  Females 1645 686 ***2.40 (2.28, 2.52) 7.00 (6.74, 7.26)  Males 1041 547 ***1.90 (1.79, 2.02) 8.61 (8.29, 8.94) 15-year mortality 1975–2020 by diagnosis year  1975–1985 479 149 ***3.19 (2.91, 3.49) 13.36 (13.05, 13.68)  1986–1995 373 137 ***2.71 (2.45, 3.00) 7.98 (7.72, 8.24)  1996–2005 492 251 ***1.96 (1.79, 2.14) 3.56 (3.35, 3.76) 15-year mortality 1975–2020 by age at diagnosis  Under 30y 114 24 ***4.76 (3.96, 5.72) 3.13 (2.99, 3.26)  30–39y 245 66 ***3.70 (3.27, 4.20) 4.79 (4.61, 4.97)  40–49y 336 126 ***2.66 (2.39, 2.96) 6.49 (6.25, 6.72)  Equal to or over 50y 649 321 ***2.01 (1.86, 2.17) 13.90 (13.50, 14.30) SMR Standardized mortality ratio, EDR excess death rate; y year *** p  < 0.001 ** p  < 0.01 * p  < 0.05

Mortality for pwMS calculated by SMR and EDR based on overall mortality during 1980–2020 for females and males, and 15-year mortality from diagnosis by both diagnosis year and age at diagnosis

SMR Standardized mortality ratio, EDR excess death rate; y year

*** p  < 0.001

** p  < 0.01

* p  < 0.05

Comorbidities were more common for pwMS deceased between 2000 and 2020 than for those alive at the end of 2020 (CCI median 1.00, Q1–Q3 0.00–3.00 vs 0.00, Q1–Q3 0.00–1.00, respectively; p  < 0.001). Overall, we saw a steady decline in SMR and increase in the age at death for both female and male pwMS from 1980 to 2020 (Fig.  2 ). Age at death steadily increased from a median of 52.0 years (Q1-Q3 43.5–60.0) before 1980, to 69.0 years (Q1–Q3 60.0–77.0) during 2010–2020, and similar findings were made for both female and male pwMS (Table  1 , Fig.  2 ). Population-based expected ages at death and ages at death of the pwMS in 10-year intervals are shown in Table  3 .

A Changes in SMR and B changes in age at death from 1980 to 2020 shown for female and male pwMS. For age at death, observed age at death is compared to population-based expected age for females (orange) and males (blue) separately

Table 3 Median life expectancies in comparison to the general population and median survival from diagnosis for female and male pwMS MS patients General population Median life expectancy (95% CI)  All patients 77.3 (76.8, 78.1) 84.0  Female 79.0 (78.5, 79.5) 85.7  Male 74.7 (73.8, 75.5) 79.0 Median survival since MS diagnosis (95% CI)  All patients 34.4 (33.3, 35.2)  Female 36.3 (34.9, 37.6)  Male 31.1 (29.2, 32.2) Median age at death (IQR) Observed Expected All patients  1980–1989 56.0 (46.0–65.0) 77.0 (72.0–79.0)  1990–1999 60.0 (50.0–70.0) 78.0 (73.0–80.0)  2000–2009 62.0 (54.0–73.0) 79.0 (75.0–81.0)  2010–2020 69.0 (60.0–77.0) 80.0 (77.0–82.0) Female  1980–1989 57.5 (46.0–66.0) 78.0 (77.0–80.0)  1990–1999 61.0 (50.0–72.0) 80.0 (78.0–81.0)  2000–2009 63.0 (54.0–75.0) 80.0 (79.0–82.0)  2010–2020 69.0 (61.0–77.0) 82.0 (80.0–84.0) Male  1980–1989 56.0 (46.0–65.0) 72.0 (70.0–74.0)  1990–1999 58.0 (48.0–68.0) 72.0 (71.0–74.0)  2000–2009 61.0 (53.0–70.0) 74.0 (72.0–76.0)  2010–2020 68.0 (59.0–76.0) 75.5 (72.0–79.0)

Median life expectancies in comparison to the general population and median survival from diagnosis for female and male pwMS

To control the follow-up time, we assessed 15-year mortality from diagnosis (Table  2 ). There was a clear decreasing trend of mortality for pwMS diagnosed between 1975 and 1985 vs those diagnosed between 1996 and 2005, with SMR decreasing from 3.19 (95% CI 2.91, 3.49) to 1.96 (1.79, 2.14), respectively. When assessed by age at diagnosis, SMR for 15-year mortality from diagnosis for pwMS diagnosed before the age of 30 years was 4.76 (95% CI 3.96, 5.72), whereas for those diagnosed after the age of 50 years, SMR was 2.01 (95% CI 1.86, 2.17).

Using life tables, we analyzed the effect of MS on life expectancy in comparison to the general population. In female pwMS, median survival was 79.0 years (95% CI 78.5, 79.5) in comparison to expected 84 years and in male pwMS 74.7 years (95% CI 73.8, 75.5) in comparison to expected 79 years. Median survival from diagnosis was 36.3 years (95% CI 34.9, 37.6) in females and 31.1 (95% CI 29.2, 32.2) in males (Table  3 ).

We then calculated hazard ratios (HR) for death according to demographic variables for pwMS. Male pwMS had a higher HR compared to females, 1.82 (95% CI 1.64, 2.00; p  < 0.001). MS diagnosis after 1995 was a protective factor from death, HR 0.55 (95% CI 0.48, 0.63) compared to earlier time intervals of diagnosis ( p  < 0.001). Older patients had obviously a higher overall HR of death compared to patients diagnosed under the age of 30 years, HR 8.58 (95% CI 7.09, 10.38; p  < 0.001) (Table  4 ).

Table 4 Hazard ratio (HR) for death according to demographic variables for pwMS Sex HR (95% CI) p value Females Ref Males 1.82 (1.64, 2.00)  < 0.001 Diagnosis year 1975–1985 1.51 (1.32, 1.74)  < 0.001 1986–1995 Ref 1996–2005 0.55 (0.48, 0.63)  < 0.001 Diagnosis age Under 30 y Ref 30–39 y 1.66 (1.34, 2.06)  < 0.001 40–49 y 2.85 (2.32, 3.50)  < 0.001 Equal to or over 50 y 8.58 (7.09, 10.38)  < 0.001 * y year

Hazard ratio (HR) for death according to demographic variables for pwMS

* y year

For the COD analysis, we collected data on COD separately during 1980–1999 and 2000–2020, and grouped CODs based on ICD-10 codes (Table  5 ). For pwMS, underlying, contributing and immediate COD data were available. Comparative data of the general Finnish population were available only for underlying COD. MS was the underlying COD in 59.1% of pwMS during 1980–1999 vs 51.3% during 2000–2020, and either underlying or contributing COD in 73.9% vs 73.5% of the patients, respectively. After MS itself, vascular diseases and cancer were the two most common underlying or contributing CODs in pwMS during both time periods. The most common immediate cause of death for pwMS was respiratory infection (23.6% during 1980–1999 and 28.0% during 2000–2020).

Table 5 Underlying, contributing COD and immediate COD in pwMS 1980–1999 and 2000–2020 Died between 1980 and 1999 ( n  = 1250) Died between 2000 and 2020 ( n  = 2686) Cause of death (COD) Underlying Underlying or contributing Immediate Underlying Underlying or contributing Immediate Multiple sclerosis 739 (59.1%) 924 (73.9%) 0 1376 (51.2%) 1973 (73.5%) 0 Other neurologic 19 (1.5%) 42 (3.4%)  < 5 118 (4.4%) 232 (8.6%) 22 (0.8%) Vascular disease 197 (15.8%) 251 (20.1%) 62 (5.0%) 424 (15.8%) 748 (27.8%) 178 (6.6%) Venous thrombosis and embolism  < 5  < 5  < 5 5 (0.2%) 8 (0.3%) 0 Chronic respiratory disease 16 (1.3%) 24 (1.9%) 6 (0.5%) 29 (1.1%) 97 (3.6%) 22 (0.8%) Respiratory infection 56 (4.5%) 71 (5.7%) 295 (23.6%) 26 (1.0%) 119 (4.4%) 753 (28.0%) Aspiration pneumonia 0  < 5 35 (2.8%) 0 10 (0.4%) 87 (3.2%) Liver disease 5 (0.4%) 10 (0.8%) 0 49 (1.8%) 77 (2.9%) 5 (0.2%) Gastrointestinal tract 19 (1.5%) 27 (2.2%) 17 (1.4%) 46 (1.7%) 91 (3.4%) 28 (1.0%) Chronic kidney and bladder disease 13 (1.0%) 31 (2.5%) 5 (0.4%) 6 (0.2%) 61 (2.3%) 17 (0.6%) Urinary tract infection 5 (0.4%) 26 (2.1%) 16 (1.3%) 8 (0.3%) 45 (1.7%) 47 (1.7%) Cancer (all types) 89 (7.1%) 98 (7.8%) 0 375 (14.0%) 440 (16.4%) 14 (0.5%) Other infection 5 (0.4%) 9 (0.7%)  < 5 9 (0.3%) 23 (0.9%) 21 (0.8%) Sepsis  < 5  < 5 36 (2.9%) 8 (0.3%) 16 (0.6%) 78 (2.9%) Musculoskeletal  < 5 5 (0.4%) 0 6 (0.2%) 29 (1.1%)  < 5 Pregnancy and childbirth 0 0 0 0 0 0 Congenital disease 0  < 5 0  < 5 9 (0.3%) 0 Disease of eyes or ears 0  < 5 0 0  < 5 0 Skin disease 0 7 (0.6%) 0  < 5 32 (1.2%) 7 (0.3%) Blood disease 0  < 5 0  < 5 24 (0.9%)  < 5 Psychiatric disease  < 5 20 (1.6%) 0 5 (0.2%) 93 (3.5%) 0 Dementia  < 5 8 (0.6%) 0 28 (1.0%) 55 (2.0%) 0 Metabolic and endocrine 5 (0.4%) 37 (3.0%)  < 5 19 (0.7%) 189 (7.0%)  < 5 Complications of medical care 0  < 5 0 0 5 (0.2%) 5 (0.2%) Suicide 17 (1.4%) 17 (1.4%) 0 46 (1.7%) 47 (1.7%) 0 Accident 19 (1.5%) 21 (1.7%) 12 (1.0%) 73 (2.7%) 93 (3.5%) 10 (0.4%) Homicide  < 5  < 5 0  < 5  < 5 0 Other 35 (2.8%) 44 (3.5%) 9 (0.7%) 16 (0.6%) 23 (0.9%) 12 (0.4%) Causes identified in < 5 patients are not reported in detail

Underlying, contributing COD and immediate COD in pwMS 1980–1999 and 2000–2020

Chronic

respiratory disease

Causes identified in < 5 patients are not reported in detail

Mortality by underlying cause (Cause specific SMR, CS-SMR; Table  6 ) during 2000–2020 was higher than expected for diseases of the gastrointestinal tract (SMR 2.15, 95% CI 1.61, 2.87; p  < 0.001), respiratory infections (SMR 1.99, 95% CI 1.35, 2.92; p  < 0.001), and for vascular diseases (SMR 1.38, 95% CI 1.26, 1.52; p  < 0.001) (Table  6 ).

Table 6 CS-SMRs in pwMS compared to the general population 2000–2020 Observed Expected CS-SMR (95% CI) Vascular disease 424 307 ***1.38 (1.26, 1.52) Cancer (all types) 375 372 1.01 (0.91, 1.11) Other neurologic 118 109 1.09 (0.91, 1.30) Liver disease 49 58 0.85 (0.64, 1.13) Gastrointestinal tract 46 21 ***2.15 (1.61, 2.87) Chronic respiratory disease 29 37 0.79 (0.55, 1.13) Dementia 28 28 1.01 (0.70, 1.46) Respiratory infection 26 13 ***1.99 (1.35, 2.92) Metabolic and endocrine 19 19 1.00 (0.64, 1.56) CS-SMR refers to underlying causes of death *** p  < 0.001 ** p  < 0.01 * p  < 0.05

CS-SMRs in pwMS compared to the general population 2000–2020

CS-SMR refers to underlying causes of death

*** p  < 0.001

** p  < 0.01

* p  < 0.05

Discussion

In this population-based nationwide study on mortality and causes of death in Finnish pwMS with up to 45 years of follow-up, we showed decreasing mortality from threefold to twofold and steadily increasing age at death through time. Median life expectancy was shortened by 7 years. Female pwMS survived a median of 36 and male pwMS a median of 31 years from diagnosis. Risk of death was lowest for the pwMS without comorbidities and diagnosis during the DMT era.

The overall SMR of 2.4 and EDR of 9 for pwMS in our cohort who died during 1980–2020 is in line with several previous studies [ 6 , 14 , 15 ]. In the previous national mortality study from Finland for pwMS diagnosed between 1964 and 1993, a threefold SMR was reported [ 10 ]. In our cohort of pwMS who died between 1980 and 1999, SMR was similarly 3.07. For patients who died between 2000 and 2020, the overall SMR had decreased to 2.14 and EDR halved from 14 to 7. Median life expectancies, years of life lost in comparison to the general population, and survival from diagnosis were similar or close to previous studies from other Nordic countries, and a recent study from the western Finland [ 6 , 12 , 23 ]. As a caveat in our study, we did not have the onset dates for the older part of our study cohort retrieved from the national care register HILMO, and thence had to use year of diagnosis instead of disease onset in the analyses.

Risk of death was significantly lower in pwMS diagnosed between 1996 and 2005 vs 1980 and 1995. Long-term studies evaluating benefit of DMTs are few and the data on effects of treatment on survival in MS are limited. Our later diagnosis period of 1995–2005 coincides with the availability of DMT’s for MS in Finland. Interferons were available first since mid-1990 s, and interferons and glatiramer acetate were the most used DMT’s until 2013 [ 21 ]. Our results indirectly suggest that lower hazard of death in our later observation period could be attributed to the availability of DMTs. Our results are in line with a previous registry-based study from western Norway, which showed that treatment-eligible pwMS diagnosed in the DMT era had the lowest risk of mortality [ 6 ].

A recent study using multiple administrative health databases from four Canadian provinces studied directly the association between the first-generation and second-generation DMTs and all-cause mortality in pwMS, followed from 1996 to 2017. They showed that earlier DMT initiation (betainterferon or glatiramer acetate vs no exposure) was associated with a significant effect on mortality, while later initiation was not. However, the survival advantage with earlier initiation diminished over time, no longer reaching statistical significance at 15 years [ 22 ]. In a recent study from Finland, concerning a regional cohort of pwMS diagnosed in 1971–2010 and followed up until the end of the year 2019, the use of betainterferon or glatiramer acetate vs no exposure was also shown to be associated with better survival, while the decade of diagnosis did not have an impact on survival [ 23 ]. In our study cohort, we did not have data of DMT use in the patients from the care register HILMO, and thence could not analyze the DMT impact on survival directly.

Alternatively, decreasing mortality in pwMS during our later observation period could be attributed to improvements in the general healthcare of the chronically disabled, such as developments in rehabilitation, symptomatic therapies and care for comorbidities, as well as diagnosis of more benign cases with increased diagnostic sensitivity. A population-based study from Denmark, spanning 6 decades, indicated that the decrease in MS mortality began well beyond the DMT era, before use of MRI became widespread, and before the McDonald diagnostic criteria were introduced [ 12 ]. They speculated that a change in the MS cohorts with fewer malignant cases may be a significant contributor. This may also be a contributing factor in our study, since we could see a decreasing SMR and increasing age at death already since 1980 s. Shorter follow-up time for the participants included toward the end of the study period also offer a plausible explanation for the improved survival.

When assessing all-cause mortality by the age at MS diagnosis, SMR was highest for the youngest diagnosis group, and lowest for the oldest diagnosis group (4.76 for those diagnosed at under 30 years vs 2.01 for those diagnosed at over 50 years). This observation was similar to previous studies from Norway and Sweden [ 6 , 16 ] and plausible as the mortality in the younger background population is low and other mortality risks increase with age. The overall excess mortality in our study was higher for females than males (SMR 2.65 vs 2.1, respectively). A previous Canadian study found a similar sex difference in mortality [ 24 ] and as in our study, but no direct explanation for this difference could be found. Speculations have included the higher incidence of MS in females, some unknown environmental factors or a difference in comorbidities, but none of these have been established. Males and those diagnosed at a higher age, however, had higher hazard ratios for 15-year mortality due to the shorter remaining life expectancy of older age groups and males. In line with this, EDRs were higher in male pwMS in our study. The opposite trends between genders in these two measures of mortality have also been found in the studies from Norway and Denmark [ 6 , 12 ]. It is considered to reflect the difference between these measures, such that EDR more directly than SMR reflects the difference in the number of lives lost [ 12 ].

Similarly, as in the previous Finnish study and several other studies [ 6 , 25 , 26 ], MS was the underlying or contributing COD in the majority of the pwMS also in the current study. Mortality by underlying cause in pwMS during 2000–2020 was higher than expected for diseases of the gastrointestinal tract (SMR 2.15, 95% CI 1.53, 2.77), respiratory infections (SMR 1.99, 95% CI 1.22, 2.75), and for vascular diseases (SMR 1.38, 95% CI 1.25, 1.51). After MS itself, vascular diseases and cancer were leading causes of death, but cancer mortality was not higher than in the general population like vascular mortality was. Infections and gastrointestinal causes were identified as causes of the excess mortality in Finnish pwMS also in the previous national study in Finland [ 10 ], and cardiovascular causes in a more recent study from Southwest Finland [ 27 ]. Suicide risk has been reported to be increased in pwMS [ 7 ], and in the previous Finnish study the SMR for suicide was 1.7 [ 10 ]. Suicide data were not available from the open data source that was used for comparison with general population in this study, and thence, SMR for suicide could not be calculated. It was the underlying or contributing cause of death in 1.7% of the pwMS in this study, a similar proportion as in a previous Swedish hospital-based study [ 28 ].

It has previously been shown that pwMS have an increased risk of developing multiple comorbidities, and comorbidities are associated with diagnostic delays and increased mortality [ 29 , 30 ]. In this study, we analyzed the impact of comorbidities on mortality using CCI as a measure of comorbidity burden. CCI predicts 10-year survival in patients with multiple comorbidities. Comorbidities were more common for pwMS deceased between 2000 and 2020 than for those alive at the end of 2020, indicating that comorbidities increased the risk of death also in our cohort. In a previous study from Finland, we observed a survival disadvantage within MS associated with comorbidity for circulatory diseases [ 27 ]. Cardiovascular diseases were among the three most common COD contributing to excess mortality also in this study. This emphasizes the importance of identification and good care of comorbidities to improve prognosis of our patients.

Strengths and limitations

The strength of our study is a large nationwide population-based cohort of over 16 000 pwMS followed up to 45 years. Limitations of the study arise from the lack of a specified control group from the general population. This limits the comparisons, such that only the underlying causes of death could be compared to the general population. Also, analysis of the effect of comorbidities was limited to pwMS, since no data of comorbidities were available for the general population. Impact of DMTs on mortality was indirectly concluded from the year of diagnosis, since DMT data were not available for patients identified from the national care register instead of the MS registry. Dates of diagnosis in the care register may have been registered with delay and were not confirmed by chart review similarly as the MS registry data is. High efficacy DMTs have only been available since last 20 years, and it is too early days to analyze their impact on mortality in a lifelong disease diagnosed at young age.

It is important to note that calculating the median age at death annually poses a limitation in the more recent diagnostic cohorts, such that many individuals may still be live at the end of the follow-up. Consequently, the observed deaths may be disproportionately from individuals with more aggressive disease courses or higher mortality risk, while individuals who have reached older ages are less likely included given the shorter observation window. Therefore, we may have underestimated the true median age at death for the more recent cohorts.

Conclusion

In summary, our study provides longitudinal population-based survival data of Finnish pwMS, demonstrating a decreasing mortality and lowered risk of death of pwMS diagnosed during the DMT era. While the mortality is still twofold and life expectancy shortened by 7 years, with reduced diagnostic delay, availability of more potent DMTs, more aggressive treatment approaches, and better care of comorbidities, the survival of MS could approach that of the general population. Future studies should address the question whether the use of high-efficacy DMTs affects mortality in pwMS.

Funding

Open Access funding provided by University of Turku (including Turku University Central Hospital). The Finnish MS Foundation, Finnish Governmental Research Funds of Southwestern Finland, Finnish Governmental Research Funds of Helsinki University Hospital.

Data availability

These data are subject to third party restrictions. Permission to access the data may be applied from Findata ( https://findata.fi/en/ ).

Declarations

Conflicts of interest

KK has nothing to disclose. SL has served as an adviser or speaker for Alexion, Argenx, Novartis, Merck, Jansen, Lundbeck, Sanofi, Teva, and UCB Pharma, and received support for congress participation from Novartis, Merck, and UCB Pharma. MV has nothing to disclose. SA has served as an adviser or speaker for Biogen, Novartis, Roche, and Sanofi. MS-H has served as an adviser or speaker for Biogen, Novartis, Roche, Sanofi, and Teva, and received support for congress participation from Biogen, Merck, and Novartis.

Footnotes

Katariina Kuutti and Sini M. Laakso had contributed equally.

References

Associated Data

Data Availability Statement

These data are subject to third party restrictions. Permission to access the data may be applied from Findata ( https://findata.fi/en/ ).