In my Swedish 28 Feburary 2016 post, I made some forecasts on the future cause of death pattern in Sweden, using age-specific death rates for ischemic heart disease, other circulatory diseases, neoplasms, and other causes for the years 1999–2013, making life tables based on the predicted death rates for future years, and based on these calculating the probability of eventually dying from the different causes, which I plotted for 2018, 2023, 2028, and 2033. According to the forecasts, the probability of dying from neoplasm (mainly cancer) would remain stable, while the probability of dying from a circulatory disease would decrease, but not so much that the ratio between circulatory disease and neoplasms would become lower than 1 during the period. This was contrary to some speculations in media that cancer would soon become the most common cause of death¹

How have these predictions stood up so far? In my 4 July 2025 post, one could see the development of the cause pattern 2013–23: the proportion of cancer deaths has not changed significantly, while circulatory diseases, as a group, have declined but are still more common than neoplasms. However, for some Swedish regions, the circulatory/neoplasm ratio is below 1, as I discussed in my 28 September 2025 post, and this has been driven by decrease in circulatory disease, and not significant increase in neoplasms.

Now, there is a difference between the observed distribution of causes of death, which is simply the ratio between number of deaths for a subset of all causes and all deaths, and the distribution given by a life table. The former depends on the actual age distribution of the population, which is influenced by past mortality, fertility, and migration patterns, while the latter, which I used for my predictions, is independent of that. This is analogous to the difference between simple average age of death and life expectancy.

I have added functions to my morr R package, to combine age-specific cause proportions, based on data from WHO (2026), with life tables in the format given by University of California, Berkeley and Max Planck Institute for Demographic Research (2026), in order to calculate life table-adjusted cause patterns, as described above. Made with these functions, fig. 1 shows the raw cause pattern for all ages for Sweden 1951–2024, while fig. 2 shows the life table-adjusted distribution of causes that a newborn eventually would die of, given the age- and cause-specific mortality rates for each year.²

Comparing fig. 1 and fig. 2, the broad patterns and trends look similar, but there are some differences. The raw ratios tend to be skewed towards the mortality patterns in somewhat younger ages, due to the factors noted above, for example higher mortality rates in the past, so that the proportion living to age 90 in life tables for 2024 is higher than that proportion in the actual 1934 birth cohort.

In 2024, 31.1 and 30.7 percent of deaths among Swedish women and men were caused by circulatory causes (the categories in the graphs down to vascular dementia), but the life table calculated ratios were 33.5/33.0 percent. In 2013, the last observation used for my 2016 prediction, the raw ratios were 39.1/37.4 percent, and the life table ratios 40.4/39.2 percent. For neoplasms, relatively more common as a cause of death in middle-age, raw ratios were 24.8/26.7 percent in 2024, and life table-based ratios 21.7/24.3 percent. In 2013, raw ratios for neoplasms were 24.0/27.2 percent, and life table-based ratios 22.3/25.6 percent.

With the life table method, both circulatory disease and neoplasms have thus decreased somewhat in Sweden since 2013, among both women and men. They have also bounced back a bit after 2020, when covid-19 entered the scene and was most common as a cause of death during the first year, but in 2024, covid mortality had become so uncommon that continued future decrease can hardly alter the pattern for other causes.

One may also note how the period with at least 50 percent circulatory causes, as discussed in my 12 april, post becomes longer using the life table method: with raw ratios, it encompasses the years 1952–94 for women and 1956–91 for men, and with life table-adjusted rations, it begins in 1951 (the first year with available statistics) for women and ends in 1998, and encompasses 1953–95 for men.

As for the future development, it is important that 2026 will be the last year using ICD-10 for mortality statistics in Sweden, according to the published schedule (Socialstyrelsen 2026). With the adoption of ICD-11, cerebrovascular diseases will be moved from the circulatory chapter to the neurological chapter, and this may well result in the circulatory/neoplasm ratio being lower 1, if the numerator is defined by the circulatory chapter. However, it is trivial to include categories from other chapters for continuity, as I have done with cerebrovascular disease before ICD-8, and with vascular dementia in ICD-10. It remains to be seen to what extent the new classification will change the cause pattern in ways harder to adjust for.

References

In my post 9 December 2024, I wrote about my Bluesky quiz, which turned out to be about time periods for which different countries had reported circulatory disease as underlying cause for at least 50 percent of deaths in their female and male populations (a property denoted X in this post).

I made the tables with some wrangling in Visidata (Pwanson 2026), but recently, I built a function in my R package morr to make similar tables over countries and years where the ratio of two causes of death exceed a certain threshold. By cloning the blog repository, and running the script 2026-04-12-half.R in the subdirectory postdata/2026-04-12-half, one should, provided that the morr package and the data files from WHO (2026) are available, be able to reproduce the tab-delimited files in postdata/2026-04-12-half/data.

These files show, for female (prop_x_f.tsv) and male (prop_x_m.tsv) populations, the lowest (column yr_min) and highest (yr_max) years, and number of years (count) where the ratio between circulatory deaths and total deaths for all ages is at lest 0.5, including only populations with at least 200 total deaths, as in the original quiz. The columns with names ending with _all show the corresponding values for years with available data for which this last condition is satisfied.

Using the latest update of WHO (2026), one can discern certain patterns, both regarding property X and the general levels of reporting during recent years of pandemic and war. Focusing on the female populations, because X has generally been more common for these, one may distinguish between the following categories.

This clearly shows what I have noted before, that X has gone from a pattern typical for rich Anglosphere or Northern European populations to something mainly seen in former Eastern Bloc countries in Eastern Europe and Asia, and has disappeared even for some of those.

It is also striking how patchy the coverage of WHO (2026) still is. Some of the largest populations, with rapid economic development in recent decades, like India and China, lack any comprehensive data. Other countries, like Syria, Ukraine, and Russia, have relatively long time-series, but have been involved in war or upheaval in recent years, and lack data from the 2020s. Modelings like IHME (2025) may be lacking in transparency, and may, to a large extent, be based on questionable extrapolations from a relatively small subset of all countries.

Both the sparsity of updated data and different practices for ascribing deaths to underlying causes complicate things like assessment of mortality related the covid-19. For many of the countries still having X in recent years, Polizzi et al. (2024) show significant life expectancy losses related to circulatory causes during 2020 and 2021, which may not be very surprising.

References

On 24 February, Statistics Sweden released new data for deaths (Statistics Sweden 2026a) as well as mean population (Statistics Sweden 2026b) in Sweden during 2025. Official life tables will be released later during March, but one can do preliminary estimates with the method i described 27 March 2023, which tend to come very close to the published ones. According to these estimates, life expectancy at birth for Sweden 2025 was 85.55 years for females, and 82.52 years for males, which implies an increase compared to the 2024 life tables, with life expectancy at 85.35 and 82.29 years for females and males (Statistics Sweden 2025). The accelerated increase after 2023 clearly reflects further decrease in mortality related to covid-19, in combination with relatively modest influenza-related excess mortality.

In my 23 December post, I wrote about the development of this influenza season in Sweden and UK so far, in the context of dire warnings about a threatening super flu, due to the drifted influenza A(H3N2) subclade K. I noted that influenza in Sweden had started rising early, but that the relative increase may have started slowing down already before Christmas, and that neither the number of cases nor the risk of severe disease, estimated by ICU cases, seemed exceptional compared to other seasons with early peak. However, I also noted that the shift to colder weather might lead to increased influenza activity the coming weeks. Fig. 1 contains an updated version of the graph in that post, based on data from WHO (2026) up to week 8.¹

As the figure shows, lab-confirmed influenza A cases peaked week 1, went down and hit a low week 4, and then started increasing again, so that the curve seemed to follow the corresponding weeks in 2025, but hit a second peak week 7, with a slight decrease week 8. A pattern with cases slowing down or decreasing around Christmas and New Year, and increasing from late January, is very common, and clearly related to changes in interpersonal contacts because of Christmas vacation, but the shift to colder weather around Christmas may have shifted the early peak slightly. Overall, the weather in Sweden was mild and humid during the first weeks of December, before the holidays, but relatively cold around Christmas and much of January and February. Had December been mostly cold and dry, we might have seen a large wave then, and relatively few cases after New Year.

For a more detailed look at the relation between influenza and holidays, it can been illuminating to study age- and sex-stratified incidence, as shown by fig. 2, which is based on data from Folkhälsomyndigheten (2026).

Cases in the age groups below 40 peaked already week 51, and the increase week 1 was driven by the oldest age group, people aged 65 or older. In that age group, contacts with younger generations may have increased during Christmas holidays. For recent weeks, cases in younger age groups have started decreasing again. This decrease may be accelerated the weeks to come, because of milder weather again, and also Winter vacation, which is during week 8 or 9 in most parts of Sweden², and will then probably spread to older age groups. Sex differences in lab-confirmed influenza within age groups are generally small, but the incidence is significantly higher among females than among males in the age group 15–39. This is, as might be expected, similar to the pattern for influenza hospitalizations, which I have written about before, for example in my Swedish 25 November 2024 post.

Cumulative ICU admissions with influenza are now at 295 for this season (Svenska intensivvårdsregistret 2026). With a continued decline in infections expected the coming weeks, it seems plausible that the season will end up at around 400 cases, like most seasons from 2015/16, except for those wholly or partially interrupted by pandemic-related contact changes.

References

During recent weeks, there has been a lot of reports, in Sweden and other countries, about a threatening severe influenza wave. In UK, media, and even health care officials, have talked about super flu (NHS England 2025). The reason for that is a fast increase in cases, driven by a new variant of the A(H3N2) subtype, the so-called subclade K (Dapat et al. 2025). This subclade has considerable antigenic drift relative to previously circulating variants of A(H3N2), including the vaccine component, and this may lead to increased \(\r\) with faster and higher spread in the general population, as well as decreased vaccine effectiveness. However, disease severity does not seem to be increased, and vaccine effectiveness against emergency department attendences and hospital admission seems to be in the normal range, at least early post-vaccination (Kirsebom et al. 2025). In England, reported cases have declined somewhat recently, from a peak lower than the seasons 2022/23 and 2024/25 (UK Health Security Agency 2025).

Clearly, Sweden has not been spared from early rise in influenza. There were reports about large rates of absence in some schools, apparently caused by influenza outbreaks, already around early Advent (Boström 2025). Fig. 1 shows reported non-sentinel influenza A cases in Sweden by week, for influenza seasons (defined by week 40 to week 20) from 2021/22 to the current season, 2025/26, i.e. all seasons since influenza returned after its absence during the covid-19 NPI:s, based on data from WHO (2026).¹

As the figure shows, influenza A cases have increased rapidly recent weeks. One may note that all the finished influenza seasons since 2021/22, except for 2024/25, have peaked around New Year, or somewhat earlier. The 2021/22 season is also probably somewhat inflated relative to the other seasons, due to the higher levels of testing, and the rapid decrease in cases during January 2022 is probably largely because of increased NPI:s due to the first omicron wave. However, 2022/23 and 2023/24 both peaked week 52, which contrasts with the most common pattern before 2020, which was more like 2024/25: exponential increase during November and December, followed by a slowdown over Christmas break, then a slower relative increase, and a February peak.

Early in the current season, most subtyped cases were A(H1N1)pdm09, but since week 47, A(H3N2) has dominated. Among the subtyped cases week 50, reported in the last WHO (2026) update, 256 were A(H3N2), while 83 were A(H1N1)pdm09. The partially interrupted 2021/22 season was also dominated by A(H3N2), while 2022/23 and 2024/25 were mixed A(H3N2) and A(H1N1)pdm09, as well as B/Victoria, and 2023/24 was dominated by A(H1N1)pdm09. Thus, Sweden has not had any uninterrupted season dominated by A(H3N2) this decade, which can be expected to contribute to increased susceptibility to this subtype, regardless of the K subclade.

The most recent week with available data for the whole country, week 50, relative increase slowed down somewhat, and the curve is now very close to 2023. Karolinska University Laboratory has published their report of respiratory pathogens with data for week 51 (Karolinska Universitetslaboratoriet 2025), and this also shows a subexponential, close to linear, increase for influenza A weeks 48–51. Maybe, the mild and humid weather in large parts of Sweden during the first weeks of December has contributed to slowing down the spread (Roussel et al. 2016). In light of that, it seems hard to make predictions for the upcoming weeks. First, \(\r\) usually decreases over Christmas, as noted above, largely because of decreased contacts during Christmas holidays. However, there has been a shift to colder and drier weather recent days, which may have the opposite effect.

As for the risk of the severe disease, it seems to be comparable to other recent seasons. There are currently 77 ICU admissions with influenza reported this season (Svenska intensivvårdsregistret 2026). This number will rise for recent days because of delay in reporting, but up to week 50, ending 14 December, there are 53 admissions. The years 2021–24, there were 35, 38, 50, and 11 admissions up to week 50. Most seasons since 2015/16 have ended up with around 400 cases, with the exception of those from 2019/20 and 2021/22 partially broken by contact changes related to covid-19. The other recent early peak seasons also fit into this pattern, with 414 cases for 2022/23, and 385 cases for 2023/24. The season with the highest number of cumulative cases is the last B/Yamagta season in 2017/18, with 498 cases, and a February peak.

So, not much evidence of any super flu so far, neither in Sweden nor UK.

References

Last week, National Board of Health and Welfare in Sweden released statistics on causes of death for the whole of 2024 (National Board of Health and Welfare 2025), which was noted in one of Sweden’s largest daily newspapers, by Nilsson (2025). That article mentions that Swedish all-cause death rates have declined with over 50 percent since 1969, which to a large extent has been driven by lower circulatory mortality. As a result of that, the article claims, cancer has become the most common cause of death in several regions. In 2024, neoplastic disease was the most common cause of death among women in Stockholm, Uppsala, and Gotland, and among men in Stockholm, Halland, and Gotland. The heading of the article says that cancer is getting more common as a cause of death.

I have written about this framing before, both in a Swedish and a US context, as in my Swedish 25 August 2016 post. Rankings of causes of death have often been based on the chapters in ICD classification. Sometimes, other partitions are used: e.g. heart disease is often treated as a category in the US. In middle- and high-income countries, circulatory disease, or heart disease, has then often been the number one cause, and neoplasms, i.e. almost entirely cancer, number two. Given this, things may seem simple: if you see a fast decline in mortality rates from circulatory disease, the ratio circulation/cancer will decrease, and eventually will be lower than 1, so that cancer will be the most common cause. And one may also then conclude that cancer must have become significantly more common in terms of its share of all deaths, even though age-standardised mortality rates still may have declined. But does that give an accurate picture of the actual development of the cause pattern in e.g. the Swedish regions?

With my morr R package, cause of death patterns over time can be plotted using the data available via WHO (2026), and also National Board of Health and Welfare (2025), for regional Swedish data, starting with 1997, i.e. the first year ICD-10 was used. Made with the functions from the morr package, fig. 1, fig. 2, fig. 3 and fig. 4 show sex-specific cause of death patterns for all ages for the four regions mentioned by Nilsson (2025).¹

As the figures show, the patterns are somewhat unstable from year to year, especially in Gotland, with its small population. But the overall pattern is clear: diseases in the circulatory chapter, i.e. the cause categories down to and including stroke (excluding vascular dementia), are getting less common, and their ratio to neoplasms has decreased, and was in 2024 lower than 1 for at least one sex in all four regions. But the neoplasm category itself has been rather stable when it comes to its share of all deaths. It is not even hard to find earlier years with somewhat higher share of neoplasm deaths for the sex- and country-combinations with a circulation/neoplasm ratio lower than 1 in 2024.

• In Stockholm, some neoplasm was underlying cause in 25.8/27.6 percent of all deaths among women/men in 2024. That share was highest among women in 2019 (26.5 percent), and among men in 2015 (28.9 percent). Already in 2005, the share was higher among men than in 2024 (27.8 percent).
• In Uppsala, some neoplasm was underlying cause in 26.8/28.8 percent of all deaths among women/men in 2024. That share was highest among women in 2021 (29.0 percent), and among men in 2011 (30.8 percent). Already in 2003, the share was higher among women than in 2024 (26.9 percent), and in 2001, it was higher among men than in 2024 (29.9 percent).
• In Gotland, some neoplasm was underlying cause in 31.3/32.2 percent of all deaths among women/men in 2024. That share was highest among men in 2004 (32.9 percent).
• In Halland, some neoplasm was underlying cause in 25.9/27.8 percent of all deaths among women/men in 2024. That share was highest among women in 2012 (28.8 percent), and among men also in 2012 (32.2 percent). Already in 2001, the share was higher among men than in 2024 (29.4 percent).

In my 4 July post, I described the overall changes in cause of death pattern in Sweden and other countries for a longer period of time, and the trends in Sweden as a whole during the ICD-10 period are, of course, also reflected on the regional level. Myocardial infarction has decreased rapidly, but there has been little change in other heart diseases, taken as a group. Stroke and arterial diseases except for ischemic heart disease and stroke (including generalized atherosclerosis) have also decreased a lot, mirroring a marked increase in dementia, which is not included in the circulatory chapter in ICD-10, even when specified as vascular. And covid-19, of course, disturbed the cause pattern when it arrived in 2020, particularly in Stockholm, where much of the spring wave was concentrated, but has more recently diminished in importance with increasing immunity in the population.

One thing that seems clear is that it is not informative to treat cause of death patterns as elections in single-member districts, where getting to number one is what matters, regardless of other aspects of the distribution of votes.

References

Last week, a newly published article about trends in US heart disease mortality (King et al. 2025) got some attention in social media. The article describes changed mortality patterns during the period 1970–2022: at the beginning of the period, heart disease mortality was dominated by ischemic heart disease (IHD), above all acute myocardial infarction (AMI). During the half-century, there has been a large decrease in overall heart disease mortality, but the decrease has been largest for AMI, so that the mortality pattern now is dominated by other types of heart disease, e.g. chronic IHD, hypertensive heart disease, and functional heart conditions, e.g. heart failure and arrhythmia. It is important to note that we are talking about underlying causes of death, which can be added because exactly one is registered for each death, and not cases where e.g. heart failure is reported as a complication of AMI.

I have written in my earlier Swedish blogs about similar patterns for Sweden, e.g. on 2 December 2011. With my morr R package, cause of death patterns over time can be plotted using the data available via WHO (2026). Unfortunately, these data cannot distinguish between AMI and other kinds of IHD for ICD versions before ICD-9, which the first countries, e.g. US, started using in 1979, and Sweden in 1987. That distinction is not possible even with recent data based on ICD-10 for some countries using the abridged List 1, e.g. Russia. Made with the functions from the morr package, fig. 1 and fig. 2 show sex-specific cause of death patterns for all ages for the US and Sweden.¹

The graphs show that the share of all deaths with heart disease as underlying cause has decreased since the 1980s, both in Sweden and the US, and that this largely has been driven by a large decrease in AMI. Non-AMI heart disease has remained relatively constant, especially in men, with some smaller variations, as with covid-19 entering the scene as a competing cause in 2020. Within that group, there has also been some drift from non-AMI IHD towards other heart disease (e.g. heart failure) and hypertensive disease. It seems rather clear that much of this is due to changed practices in certification, coding and selecting underlying cause, e.g. the shift to ICD-9, which was accompanied by an increase in other heart diseases, both in the US and Sweden. The remarkably low share of deaths due to hypertensive disease in Sweden in the early 1980s points to a strong tendency to attribute cardiac deaths, where hypertension may have been involved, to IHD.

It is important to note that there will soon not be much room left for further lowering the share of all deaths due to heart disease by decreasing AMI mortality, because the share of heart disease deaths due to AMI will be so small that further decrease will not make much difference. And the decreasing share of AMI deaths may also have consequences for the validity of using models based on older data for population-level attribution of shares of IHD, or heart disease, mortality to risk factors like cholesterol, which may largely influence mortality through their effects on conditions closely related to AMI (Menotti and Puddu 2019).

For the overall decrease in the share of deaths due to circulatory disease, as defined in ICD-8 to ICD-10, on which I made a quiz, as I wrote about in my 9 December 2024 post, the decrease in cerebrovascular disease has also been significant. However, I have plotted that group next to vascular dementia and other dementia-related categories (including e.g. Alzheimer disease and unspecified dementia). If one looks at these as a group, the changes have not been that dramatic. And that would probably have been the default way at looking at them, had cerebrovascular disease not been moved from the chapter for neurological disease to circulatory disease in ICD-8. With ICD-11, the category will be moved back, and it may again seem natural to look at age-related neurological disease as a spectrum, from more acute strokes over vascular dementia to mixed, Alzheimer or unspecified dementia, similar to how the remains of the circulatory chapter group AMI with chronic IHD, as well as functional and hypertensive heart disease. One may also note that the group of arterial disease except for IHD and stroke has decreased in parallel with the increase in dementia. This group includes conditions like generalized atherosclerois (ICD-10 I70.9), which may be regarded as garbage codes, when used for underlying causes of death (Naghavi et al. 2010), and may in the past often have been used in patients with dementia without other obvious causes of death.

As noted in my December post, there are still countries with more than 50 percent of deaths due to circulatory disease, and many of these are former Eastern Bloc countries. However, we should not, in that context, regard these as simply present-day versions of Northern European and Anglosphere high-AMI countries during late 20th century. As I noted in my 19 July 2022 post, their life expectancy may be as high as some Western countries like US in recent years, at least among women. And the pattern within the circulatory categories may also be very different from the older Western pattern. As an example, fig. 3, shows the cause distribution for Lithuania, for the same categories as in fig. 1 and fig. 2.

As can be seen, Lithuania has a low proportion of deaths due to dementia, and a high proportion due to circulatory disease, like the US and Sweden 50 years ago. But there are also striking differences, with the IHD category being larger among women than among men in Lithuania, and AMI only consisting a small fraction of IHD, regardless of time period. This is in line with the findings by Timonin et al. (2021), discussing differences in ascribing IHD deaths to AMI between Norway and Russia (using detailed Russian data not available via WHO (2026), as noted above). It remains to be seen whether future development, e.g. the adoption of ICD-11, will lead to greater convergence between Western and Eastern countries regarding these patterns.

References

Around New Year, the Swedish newspaper DN had an article about the spread of contagious diseases, like influenza and covid-19 (Letmark 2024). According to the article, the viruses are spreading rapidly because we have lots of contacts around Christmas and New Year. Sweden has indeed had to influenza seasons in a row with reporting cases peaking in last week of the year. At the same time, one must remember that this is due to gradual build-up of infections the weeks leading up to Christmas, when people have lots of contacts at schools and workplaces. If cases peak around New Year, this implies that \(\r\) is on its way down. As the article also states, the spread often increases after start of spring semester, but if a lot of immunity has been built up, we may avoid later peaks.

In my Swedish 25 November post, I wrote about infectious disease trends in Sweden during fall, with covid-19 on a plateau, Mycoplasma pnemumoniae infections at high levels, and flu still at low levels. Updated reports around the holidays are sparse, but data up until week 52 are available for Stockholm (Karolinska Universitetslaboratoriet 2025). Not much has happened with covid-19, and M. pnemumoniae infections are now on their way down. Influenza A has increased during December, but is still at lower levels than during the December months 2021–23, and the increase has stalled during week 52. As noted, we have yet to see whether there will be large increases after the holidays.

An important reason for the early influenza peeks during 2021–23 is most certainly that population immunity has been lower than usual, because influenza almost stopped circulating during the pandemic. Now, when we have had both A(H1N1)pdm09, A(H3N2) and B/Victoria circulating after 2021, we may see a return to a pattern with influenza peaking after New Year. Such a pattern for influenza peaks has been most common in Sweden for several decades, or more. In my Swedish 22 June 2020 post, using data back to 1981, I noted that peaks in total mortality already in December has coincided with severe waves of seasonal influenza, as in 1988 and 1993.

On one of my old blogs, I wrote a Swedish post 20 January 2011 about cause-specific mortality in Sweden per month, for the years 1969–86, when ICD-8 was used and monthly statistics on underlying causes of death in Sweden were included in published reports, later made available via SCB (2019). For the ICD-10 years, 1997 and onwards, such statistics has recently been made available via National Board of Health and Welfare (2024). For the ICD-8 period, the monthly statistics is available for all ages, and ages above 75, but for the ICD-10 period, only statistics for all ages is currently published.

In the monthly ICD-8 statistics, important categories for respiratory infections include acute respiratory infections and influenza (ICD-8 460–478) and pneumonia (ICD-8 480–486), while the newer statistics uses the block influenza and pneumonia (ICD-10 J09–J18), where the use of the and conjunction often causes confusion, both in Swedish¹ and English as it might be interpreted as referring to having both influenza and pneumonia, not (at least) one of the diseases. Fig. 1 and fig. 2 show number of deaths among females and males in Sweden, all ages, per month 1969–86, for the mentioned ICD-8 categories, and fig. 3 shows these data for influenza and pneumonia 1997–2024, with preliminary data available up until September 2024.²

The graphs clearly show that mortality peaks already in December or January have been relatively uncommon during both periods. In 1969, the A(H3N2) pandemic peaked around New Year, and caused significant peaks in both of the respiratory categories. After a calm 1970/71 season, there was a new peak in late 1971, which perhaps may be seen as a second pandemic wave in Sweden. Otherwise, there are no clear December peaks during the 1969–86 period. The severe A(H3N2) seasons 1988/89 and 1993/94 occurred both during the ICD-9 period, for which no monthly cause-specific statistics has been published yet, even though the excess total mortality is clear, as noted above. For the ICD-10 period, December and January peaks in the respiratory categories can be seen for 1999/2000, 2003/04, 2016/17, and the latest years, from 2021/22. Note that the latest seasons are not large in size, compared to e.g. 2018. Even though immunity debt after the decreased circulation in 2020–21 seems to have influenced the timing of the influenza seasons, they have not necessarily been much larger than usual in terms of e.g. mortality. In this respect, antigenic drift, relative to the strains people have been exposed to earlier in life, may be more important than waning immunity for influenza (Munro and House 2024).

One must remember that these respiratory captures by no means capture all excess mortality caused by influenza. The smaller ICD-8 category, acute respiratory infections and influenza is probably rather specific for influenza-related mortality, but has low sensitivity. The categories including pneumonia are probably neither very sensitive or specific. As I noted in my 2011 post, much influenza-related mortality is ascribed to circulatory causes, and in recent years probably often dementia. For example, fig. 4 shows deaths among females and males in Sweden 1969–86 with other ischemic heart disease (ICD-8 411–414) as cause of death.

There are clear peaks in this type of mortality coinciding with, for example, the flu peaks during March 1976 and March 1983. One may note that the category is rather diffuse, and that many of these deaths in recent years probably would have been ascribed to other causes, such as dementia, in accordance with the trends I discussed in my 9 December 2024 post.

For several seasons during the late 1970s and early 1980s, there is very little excess mortality from influenza. This coincides with the reintroduction of A(H1N1) 1977, where older generations were protected by immunity they had acquired when this subtype circulated before 1957. For similar reasons, they 2009 A(H1N1)pdm09 pandemic did not cause much excess mortality either. There are also seasons dominated by influenza B, with low average age among cases, and similarly low excess mortality. One may take the 1983/84 as an example. During the winter, Moscow reportedly had high influenza activity, mostly influenza B, and most cases occurring among children and young adults (CDC 1984). Eventually, the epidemic reached Sweden, and newspaper articles, searchable via Kungliga biblioteket (2024), indicate a peak in Stockholm in early April, while mortality was at low levels.³ In contrast, the relatively high mortality during the possibly last B/Yamagata season, 2017/18, when cases were much older on average, as I wrote in my Swedish 8 February 2020 post, shows that this depends on population immunity, and not on influenza B being more benign by itself.

References

Nu till nyår hade DN en artikel om spridning av smittor som influensa och covid-19 (Letmark 2024). Det påstår att virus sprids med hög fart, därför att vi träffas och umgås med varandra ganska tätt vid jul och nyår. Ja, vi har nu haft två influensasäsonger i rad där de rapporterade fallen toppade just årets sista vecka. Men det är viktigt att komma ihåg att detta beror på att smittan byggts upp veckorna innan, när folk träffats på skolor och arbetsplatser. Om de rapporterade fallen toppar kring nyår, innebär det att \(\r\) är på väg ned. Som också påtalas i artikeln tenderar smittspridningen sedan ofta att öka efter vårterminens start, men om mycket immunitet hunnit byggas upp före nyår, kan det hända att det inte blir någon senare topp.

Den 25 november skrev jag litet om utvecklingen under hösten, där covid-19 hamnat på en platå, mykoplasma i luftvägarna nått höga nivåer och influensan ännu inte tagit fart. Mycket av rapporteringen ligger nu nere över helgerna, men data till och med vecka 52 finns tillgängliga för Stockholm (Karolinska Universitetslaboratoriet 2025). Det är fortsatt mycket små förändringar när det gäller covid-19, och infektioner av M. pneumoniae är definitivt på väg nedåt. Influensa A har ökat under december, men den befinner sig på lägre nivåer än decembermånaderna 2021–23, och ökningen har bromsats upp vecka 52. Som sagt återstår att se vad som händer när vårterminen börjar.

En viktig anledning till de tidiga influensatopparna de senaste åren är säkerligen att immuniteten i befolkningen varit lägre än vanligt under hösten efter att cirkulationen av influensa hållits nere under pandemin. När nu både A(H1N1)pdm09, A(H3N2) och B/Victoria cirkulerat efter 2021, kanske det innebär en återgång till ett mönster där influensan oftast toppar efter nyår. Ett sådant mönster för influnesatopparna har varit det vanliga i Sverige under lång tid. Som jag visade på den 22 juni 2020, med data tillbaka till 1981, har dödlighetstoppar redan i december sammanfallit med svåra säsongsinfluensor, som 1988 och 1993.

Den 20 januari 2011 skrev jag på min då aktiva blogg om orsaksspecifik dödlighet i Sverige per månad, för åren 1969–86, då ICD-8 användes och månadsstatistik över underliggande dödsorsaker i Sverige publicerades, som sedan gjorts tillgänglig via SCB (2019). För perioden med ICD-10, från 1997 och framåt, finns liknande statistik numera tillgänglig via Socialstyrelsen (2024). Statistiken för den äldre perioden finns publicerad för alla åldrar och åldrar från 75 år, medan den för den nyare perioden endast finns för alla åldrar.

För månadsstatistiken baserad på ICD-8 finns för luftvägsinfektioner kategorierna akuta luftvägsinfektioner och influensa (ICD-8 460–478) och lunginflammation (ICD-8 480–486), men för den nyare statistiken finns den sammanslagna influensa och lunginflammation (ICD-10 J09–J18), en benämning som flera gånger gett upphov till förvirring, vilket jag skrev om den 14 mars 2020. Fig. 1 och fig. 2 visar döda kvinnor och män i Sverige per månad 1969–86 för de nämnda luftvägskategorierna från ICD-8, och fig. 3 visar motsvarande data för influensa och lunginflammation 1997–2024, med preliminär statistik till september 2024.¹

Det är tydligt att dödlighetstoppar redan i december eller januari varit ganska ovanliga under båda perioderna. År 1969 kom pandemin med A(H3N2), och den toppade kring nyår, vilket syns på båda luftvägskategorierna. Efter en lugn säsong 1970/71 blev det på nytt en topp i slutet av 1971, som möjligen skulle kunna betecknas som en andra pandemivåg i Sverige. I övrigt är det svårt att se några klara decembertoppar under hela perioden. De svåra A(H3N2)-säsongerna 1988/89 och 1993/94 inträffade alltså under perioden då månadsstatistik över dödsorsaker inte publicerats, även om statistiken över total dödlighet talar sitt tydliga språk, som jag visat på i tidigare inlägg. För perioden med ICD-10 går det att se december- eller januaritoppar i influensa och lunginflammation kring millennieskiftet, 2003/04, 2016/17, och de senaste åren, från 2021/22.

En viktig sak är att dessa luftvägskategorier inte fångar upp all överdödlighet relaterad till influensa. Den mindre kategorin baserad på ICD-8 är troligen relativt specifik för influensa, men inte speciellt sensitiv. Kategorierna för lunginflammation, eller influensa och lunginflammation är varken speciellt sensitiva eller specifika. Som jag visade på i mitt inlägg från 2011 är det en hel del influensarelaterad dödlighet som tillskrivs hjärtsjukdomar, och på senare år säkerligen demens. Som ett exempel visar fig. 4 nedan dödsfall bland kvinnor och män 75 år och äldre i Sverige 1969–86 med annan ischemisk hjärtsjukdom (ICD-8 411–414) som dödsorsak.

Här syns t.ex. tydliga toppar som sammanfaller med influensatopparna marsmånaderna 1976 och 1983. Det kan noteras att detta också är en ganska diffus kategori, där många dödsfall i dag troligen hade tillskrivits t.ex. demens, i enlighet med de trender jag senast tog upp i inlägget den 9 december förra året.

Det finns också flera säsonger under sent 1970-tal och tidigt 1980-tal där det tycks ha varit mycket litet influensarelaterad dödlighet. Detta sammanfaller med återintroduktionen av A(H1N1) 1977, där äldre generationer skyddades av immunitet som de förvärvat när subtypen cirkulerade före 1957. Av liknande skäl gav inte heller pandemin 2009, med A(H1N1)pdm09 upphov till någon tydlig topp i dödlighet. Även säsonger dominerade av influensa B, som också haft låg medelålder bland fallen och medfört låg dödlighet, har förekommit. Ett exempel kan vara 1983/84. I mars 1984 rapporterades att det varit hög influensaktivitet i Moskva, primärt kopplad till influensa B, och att de flesta drabbade var barn och unga vuxna (CDC 1984). Den nådde tydligen också Sverige, och artiklar som kan sökas via Kungliga biblioteket (2024) tyder på en topp i Stockholm under första halvan av april, samtidigt som dödligheten alltså var på låga nivåer.² De relativt höga dödstalen den kanske sista säsongen med B/Yamagata 2017/18, när de drabbade var äldre, som jag skrev om den 8 februari 2020, visar att detta beror på immuniteten i befolkningen, och inte på att influensa B skulle vara lindrigare i sig.

Referenser

On 26 November, i posted a little quiz on Bluesky, with screenshots of Visidata (Pwanson 2026) showing year intervals during which a certain epidemiological property held for female and male populations (Pettersson 2024). The tables are also available in the blog repository as tab-delimited files in the directory postdata/2024-12-09-quiz. I asked about the property they show.

The answer is that they show the first and last, or latest, year, and total number of years, with mortality statistics available via WHO (2026), for which at least 50 percent of deaths among females or males, over all ages, have circulatory disease as underlying cause of death. To avoid fluctuations caused by small populations, in microstates and other territories in the WHO statistics, only years with at least 200 total deaths for each sex have been included.

Circulatory deaths are defined according to the expressions in the Mortintl.jl counfiguration, i.e. mostly the circulatory chapter in the different ICD versions, plus cerebrovascular diseases, which were included in the neurological chapter in ICD-6/7, as they will be again in ICD-11 (WHO 2024).

The tables show patterns similar to the ones I discussed earlier, e.g. in my posts 27 March 2022 and 19 July 2022. In general, having at least 50 percent deaths ascribed to circulatory disease, has been more common among females, despite females having lower risk for early circulatory death than males in most populations.

The earliest statistics available via WHO (2026) is from 1950, and at that time, populations in the Anglosphere, such as US, Australia, New Zealand, England and Wales and Scotland, had the property for at least their female populations, and Nordic countries soon followed. However, in most of these countries, the period ended in the 1980s. Sweden was relatively late, with 1991 and 1994 as last years for males and females. The last European non ex-Communist countries having it for at least their female populations were Austria and Germany (partially former East Germany, however) in 2003, and Greece in 2008. Most of the countries with end years later than 2000 are former Eastern Bloc countries in Eastern Europe or Central Asia, and some Middle East and North African countries, such as Syria and Egypt. It should be noted that the availability of statistics via WHO (2026) is very sparse for many countries, especially in Africa.

These developments could be interpreted partially as a story of rising and falling mortality from circulatory diseases, due to medical advances and environmental factors. But, as I have touched on in my earlier posts, they probably, above all, tell us a story about changing views of what we accept as causes of death, as discussed by Reid (2024). Rich countries in the Anglosphere were quick to largely dismiss causes like old age during the 20th century, which led to circulatory diseases, often referring to arteriosclerosis, dominating age-related mortality instead (Kohn 1982). But later, these countries, with their ageing populations, were also quick in perceiving also these practices as too rough, which has led to increasing shares of deaths attributed to other causes, such as dementia. Thus, the typical Western late 20th century pattern, with a very high proportion of circulatory deaths, is nowadays most common in middle-income countries. The mentioned changes in ICD-11, where stroke and other cerebrovascular diseases are separated from heart disease and merged together with dementia, may be seen in light of these changed perceptions.

References

Den 31 mars i år skrev jag om hur sjukhusinläggningar för covid-19 i Sverige gradvis närmat sig motsvarande tal för influensa, framför allt i yngre åldersgrupper. När jag skrev detta hade jag tillgång till preliminär statistik för 2023 och ingen alls för 2024, via Socialstyrelsen (2024). Nu har preliminär statistik till och med augusti 2024 publicerats, med vars hjälp jag skapat uppdaterade versioner av graferna i marsinlägget.

Som i det inlägget visar fig. 1 utskrivna från slutenvård i förhållande till folkmängden för covid-19 (ICD-10 U07–U09) som huvuddiagnos bland kvinnor och män i 5-åriga åldersintervall, baserat på data från Socialstyrelsen (2024) och SCB (2024), över varje säsong (från juli ett år till juni följande år) då covid-19 cirkulerat. Fig. 2 visar motsvarande incidens för influensa (ICD-10 J09–J11) tillbaka till 1998, då statistiken för slutenvård börjar. Fig. 3 visar incidens i covid-19 relativt högsta incidens i influensa för någon säsong under perioden.¹

Som väntat, sett till att covidfallen toppade före nyår, syns inga dramatiska förändringar av covidkurvan. För alla åldersgrupper var andelen sjukhusvårdade lägre 2023/24 än 2022/23, och för åldersgrupperna från 15–19 till 70–74 år var det också den lägsta säsongen under hela den tid SARS-CoV-2 cirkulerat. Överrisken för män när det gäller sjukhusvård har försvunnit i åldersgrupperna under 60 år, men finns kvar i äldre åldersgrupper och är störst bland de äldsta.

Kurvorna för influensa för säsongen 2023/24 har förskjutits uppåt, vilket också är väntat, då en stor andel av fallen rapporterades efter nyår. Säsongen som helhet ter sig fortfarande helt ordinär jämfört med prepandemiska influensasäsonger.

Vad kan vi vänta oss av den kommande vintern? Under sommaren och hösten har det varit, för säsongen, relativt mycket spridning av SARS-CoV-2, med ett mönster som litet liknar den sega platå vi såg 2022, och som jag diskuterade den 18 mars 2023. De senaste veckorna har det dock varit en långsam minskning av rapporterade fall, som inte setts under oktober och november något tidigare år sedan pandemin började (Folkhälsomyndigheten 2024). Karolinska Universitetslaboratoriet (2024) visar på minskning även för vecka 47. Som visas av Hodcroft (2024) har påtagliga skiften när det gäller varianter av SARS-CoV-2, som ökat virusets förmåga att undkomma befolkningens immunitet mot smittsam infektion, blivit mer sällsynta jämfört med den första tiden efter introduktionen av omikron 2021. Sedan våren 2022 har det i stort sett enbart varit subvarianter av BA.2 som cirkulerat. Under andra halvåret 2022 dominerade BA.5 och dess avkomling BQ.1, men den kommande vintern tog XBB, en rekombination av två andra subvarianter på BA.2-trädet, över scenen, för att dominera under större delen av 2023, med nya subvarianter och rekombinationer. I mitten av året kom BA.2.86, som är ytterligare en subvariant av BA.2, utan närmare släktskap med de ovannämnda, och den blev i Sverige dominerande mot slutet av året. Den har fortsatt att utvecklas under 2024, men inga nya skiften till avlägsna subvarianter har observerats. Vi har troligen nu ganska omfattande immunitet mot BA.2.86 i befolkningen, och därmed verkar det sannolikt att den kommande vintern blir än lugnare än den förra när det gäller covid-19.

Influensan har inte tagit fart så långt vi kan observera, även om den senaste veckans kyla kan komma att visa sig i ökat antal fall framöver. Sedan influensan återkom efter uppehållet under pandemin har alla tre varianterna, A(H1N1)pdm09, A(H3N2) och B/Victoria, hunnit cirkulera och dominera i omgångar, vilket innebär att vi troligen har ett ganska gott immunitetsläge även här. B/Yamagata tros av många ha försvunnit 2020, vilket kan hänga samman med att den cirkulerade på låga nivåer redan innan de pandemirelaterade kontaktminskningarna började, även om det för några veckor sedan rapporterades ett fall från Nederländerna, vars status fortfarande är oklar (ECDC 2024). I övrigt har diskussionen kring influensa den senaste tiden varit fokuserad på pandemihotet från A(H5N1).

Något som fått en hel del uppmärksamhet i media under hösten är luftvägsinfektioner orsakade av Mycoplasma pneumoniae. Av Karolinska Universitetslaboratoriet (2024) framgår att de rapporterat fallen ökat kraftigt sedan i somras, efter att smittan varit borta under pandemiåren, börjat komma tillbaka till sommaren 2023 och sedan i stort sett legat på en platå över vintersäsongen 2023/24. Möjligen kan den nu ha kulminerat. Infektionen är inte anmälningspliktig, och det finns ingen rikstäckande statistik publicerad. Studier av sjukhusinläggningar till följd av M. pneumoniae med hjälp av statistik publicerad via Socialstyrelsen (2024) försvåras också av att dessa enbart redovisar ICD-koder för huvuddiagnos på tredjepositionsnivå, samtidigt som lunginflammation orsakad av M. pneumoniae är en fjärdepositionskod i ICD-10 (J15.7). Det verkar dock troligt att reducerad immunitet till följd av att smittan hållits nere under pandemin bidragit till ökad spridning det senaste året. En liknande ovanlig ökning av infektionen observerades under 2011, vilket jag skrev om den 1 november 2012. Kanske kan även de relativt små förändringar i kontaktmönster som ägde rum under pandemin 2009 ha varit en bidragande faktor som rubbade dess normala mönster. Om det dessutom är så att den, till skillnad från många virus, sprids lättast under sommar och tidig höst, skulle det möjligen kunna förklara varför den inte tog fart redan förra vintern.

Referenser