When swine flu poked its head above water in the northern hemisphere in April our “normal” flu season was just ending. A surge of swine flu cases during a time when influenza was not usually seen was bewildering and confusing, not to mention alarming. We didn’t know what to expect nor were we sure if it would peter out over the summer, as flu usually seems to (we may learn differently in the future as we start to do surveillance over that period) and then come roaring back or just disappear. While it didn’t peter out very much, the big question was going to happen when the northern hemisphere entered its traditional flu season when, for reasons we still don’t understand, influenza infections greatly increase. CDC’s official flu season has been week 40 of one calendar year (October) to week 20 of the next (April). But this was a pandemic strain so we didn’t know how it would act when it had to compete with seasonal flu viruses under environmental conditions that seem to favor infection. So almost immediately eyes turned to the southern hemisphere which was on the cusp of its usual flu season (6 months out of phase with ours). We still don’t have all the results but a paper published this week in the New England Journal of Medicine summarizes experience with critical care bed demand in Australia and New Zealand.
The demand for critical care beds (“Intensive Care Units”) is one of the nasty problms in an increasingly brittle health care delivery system in the US. As noted in an earlier post, it is quite common for big city hospitals to have periods where they are “on diversion,” meaning that ambulances carrying patients to their emergency departments (ED) are diverted to another hospital, not because of lack of space in the ED but because of lack of staffed critical care beds. I am told much of this may be the result of stupid scheduling for elective surgeries (not spreading them evenly throughout the week), but in general ICU beds, especially those with mechanical ventilation, are the neck of the acute care hourglass. So it’s useful to make some back-of-the envelope calculations to see what might happen:
We performed a multicenter inception-cohort study involving 187 ICUs in Australia and New Zealand ? all the ICUs (adult, pediatric, or adult and pediatric) in the two countries.12 The ICUs had a total of 1879 beds, of which 1449 were equipped for mechanical ventilation.. . .
From June 1 through August 31, 2009, we identified all patients admitted to the ICU with confirmed infection with the 2009 pandemic influenza A (H1N1) virus. The 2009 H1N1 influenza was confirmed by means of a polymerase-chain-reaction (PCR) assay or serologic analysis. The 2009 pandemic influenza A (H1N1) virus and seasonal subtypes (preexisting H1N1 and H3N2 strains) were confirmed by PCR assay. (ANZIC Influenza Investigators, NEJM)
The investigative team obtained data for each patient on dates and times of admission and first symptoms, medical condition on admission (especially airway status), age, sex, race, BMI [Body Mass Index], pregnancy or childbirth less than 28 days previously and a large number of pre-existing medical conditions. 856 patients with laboratory confirmed influenza A infection wound up in an Australian or New Zealand ICU in the 3 month period, 722 (84.3%; 626 Australia, 96 New Zealand) with identified swine flu. Most of the rest were not able to be typed further and it is likely many or perhaps most were also swine flu cases. 37 patients (4.3%) had infections with identified seasonal flu subtypes (seasonal H1N1 or H3N2), so of the influenza patients in critical care beds, the overwhelming proportion were swine flu cases. Comparing this number (722) to previous years for patients admitted to the two countries’ ICUs for the general category of viral pneumonitis (which includes seasonal influenza A) the numbers are 57 in 20º5, 33 in 20º6, 69 in 2007 and 69 in 2008 (mean of 57 patients for the same 3 month period that yielded 722 this year; the seasonal flu number of 37 is on a par with other years). So the admission rate for influenza is at least 15 times higher than in other years. It still represents a minority of ICU admissions, however.
The ICU admission rate for Australia and New Zealand (combined population of roughly 25 million) was calculated as about 30 per million (I have rounded slightly because these are estimates). In the US, with a population of 300 million, this would amount to about 9000 ICU bed admissions for swine flu over a comparable 3 month period (the flu season isn’t over in the southern hemisphere but the objective was to get some preliminary data that would be of use to planners in the north).
How does this compare with the US? According to Table 3 in this paper (subscription only, I’m afraid) in 1999 the US had 5.7 million adult ICU admissions/yr. (so this underestimates a bit because pediatric ICUs aren’t included but it’s likely we have fewer beds per population than before as well; it’s the best data I could find). So that’s roughly 1.3 ICU million admissions in a three month period. If swine flu contributes another 9000, that’s not much added burden (maybe 1%) and even in terms of bed days it’s only about 2% of the total bed days, although it will be very unevenly distributed, with some places hit hard and others barely touched. As usual with flu, the paper showed a lot of geographic variation. It also showed usage peaking 4 to 6 weeks after the first winter admissions for swine flu, so we have yet to feel the bite in critical care beds. It’s also important to remember that beds don’t care for patients, nurses and doctors do, and staffing shortages is another area where we have been having trouble. I expect that we’ll be seeing some trouble spots, but the data as we see them here don’t suggest a looming catastrophe. We’ll find out soon enough.
The striking epidemiology of swine flu seen elsewhere is again confirmed in these data. The two graphs below show rates and numbers for different age groups. The top panel gives rates, and the highest risks are seen in the 0 to 1 age broup (babies), followed by adults in the 25 to 64 year old age ranges. But because there are so many more people in the 25 to 49 year olds, the numbers of ICU admissions in that age group is highest (bottom panel). Consistent with the initial stages, risks in the over 65 group are relatively low (as are number of ICU admissions), unlike seasonal influenza where they represent most of the severely ill flu cases.
Major risk factors are a BMI over 35 (CDC considers 30 to be obese; a five foot six inch female with a BMI of 35 would weigh 216 pounds and a 5 foot 10 inch man 243 pounds; table here). Almost a third of the ICU admissions had a BMI over 35, a level of obesity found in only 5% of the population down there. During the same period it was estimated that about 1% of the population in those two countries was pregnant, but pregnant women made up over 9% of the ICU admissions, once again confirming pregnancy as a major risk factor. Among other pre-existing medical conditions, asthma and chronic pulmonary conditions were present in about a third of the ICU admissions but estimated to be only 13% prevalence in the general population. Being a member of one of the indigenous populations was also a risk factor, something seen in Canada’s First Nation and eskimo communities as well. But about a third of the ICU admissions were in healthy adults with no known predisposing factors.
As might be expected with people admitted to ICUs, these folks were pretty sick. About half had Acute Respiratory Distress Syndrome (ARDS) or viral pneumonia and 20% had a secondary bacterial pneumonia. Once again, if the thought is that modern medicine will save us because we have antibiotics and respirators, these data don’t suggest this will make a major impact. Interestingly, the proportion who died (about 16%) is essentially the same as people who were admitted to the ICU for regular seasonal flu subtypes. The main difference is in the ages: with swine flu it’s people below the age of 65, with seasonal flu it’s people over the age of 65.
We don’t know how much relevance these data will have for countries in the northern hemisphere, where the availability of critical care beds varies dramatically among developed countries as do medical care practices. Complicating this is the recommendation if not requirement that patients with swine flu be isolated, meaning that the usual multibedded ICU may be unsuitable, putting added stress on resources. Since total increase in demand is only a few percent, if it is not too sharp and too high it should be possible to make adjustments by converting acute care beds to critical care beds and postponing elective surgery. Hospitals that have done some pandemic planning will be much better off than those that haven’t. Unfortunately many have only paid lip service to detailed planning. They hospitals will struggle if hit suddenly.
The big unknown is what effect the timing, availability and uptake of swine flu vaccine will have. There was no vaccine for swine flu throughout the time reported by this paper. No one has any good idea what will happen in the northern countries now that flu season is starting. I’m taking out a public health insurance policy and getting swine flu vaccine (I already got the seasonal flu shot) as soon as my turn comes up, even though I am in the lower risk age group. The risk is still substantial and it’s the most rational thing to do.
I try to be rational about things, although I’m not always successful. This one I’ll make sure I succeed at.