It is still too soon to say that we understand how Swine Flu kills people and how often this happens.
Ostensibly the proportion of infections that kill people is higher in Mexico than in the US and other affected countries. However, this could be misleading, because it is very likely that in Mexico, only the most severely affected cases would have come to the attention of the authorities. In contrast, in the USA, with greater awareness and better health services, the authorities could have been made aware of milder cases, and hence the death rate would seem to be low, as reported.
More work is needed to ascertain more (hopefully all) of symptomatic persons, so that the risk of death can be better quantified. 20-70% of infections may in fact be asymptomatic, which explains how the virus can spread by sometimes “invisible” chains of transmission.
As far as how the deaths occur — it is very likely that early deaths (i.e. within 1-4 days of infection) are at least partly related to the cytokine storm (inflammatory response), triggered by influenza virus interacting with T-cells and macrophages in the absence of protective antibody. Later deaths are more likely to be due to complicating bacterial infection, but many of these should be treatable by antibiotics, provided that the lung damage caused by the virus and preceding inflammation is not too severe.
The severity of symptoms, and the risk of death is also related to age and population characteristics. In the 1918-19 pandemic, children in isolated Alaskan villages survived, while most of their parents and grandparents died. The high death rate in adults in these isolated populations was probably because they had never (or rarely) been visited by even seasonal forms of influenza and hence lacked protective antibody to any form of influenza. Children were probably protected by innate immunity, which seems to have faded by adult life. Another possibility is that by adult life, T-cells have become more switched-on, making the cytokine storm worse in adults in the absence of protective antibody.
In contrast, the mortality pattern in urbanised populations in 1918-19, was somewhat different, in that older adults (born before about 1890) had a lower death rate than younger adults, arguably because they still had protective antibody from a related virus that had circulated prior to 1890 and then disappeared. Furthermore, the influenza death rate at all ages in urbanised populations was much less than in those immunologically naïve Alaskan villages.