Oxygen therapy in swine flu

A study published today in the Journal of the American Medical Assocation has looked at the outcomes of people with severe respiratory complications from swine flu who required a specific treatment to oxygenate their blood.

The study was in all people with swine-flu associated acute respiratory distress syndrome (ARDS) in Australian and New Zealand intensive care units during the winter flu season. Patients were given extracorporeal membrane oxygenation (ECMO). This involves pumping a patient's blood through a machine that oxygenates it and removes carbon dioxide before pumping it back into the body.

By the end of the study, 71% of patients had been successfully discharged from intensive care, 9% were still in intensive care, and about a fifth died (21%). The researchers say that this is a relatively low death rate considering the severity of their condition and the intensity of the ECMO treatment.

Such studies are useful for planning purposes. The researchers estimate that the EU may need to provide ECMO to approximately 1,300 patients during this winter season.

Key points

  • This study looked at the characteristics and outcomes of all 68 intensive care unit (ICU) patients in Australia and New Zealand who received ECMO for ARDS associated with confirmed or suspected swine flu between June and August 2009 (their winter flu season).
  • This represented 2.6 people out of every million in the population who needed ECMO for ARDS associated with confirmed or suspected swine flu, compared with an estimated 0.15 cases per million in the previous flu season. The proportion of people with swine flu requiring ECMO was not calculated.
  • Based on these figures the study’s authors predicted that the EU might need to provide ECMO to about 1,300 patients during the upcoming winter flu season.
  • Patients who needed ECMO were mainly young adults (average [median] age of 34.4), with only three children aged under 15 and no elderly patients requiring ECMO. About half of the patients were obese, 28% had asthma, 15% had diabetes, and 15% were pregnant or had recently given birth.
  • On average, ECMO continued for 10 days (median), median length of ICU admission was 27 days, and the median length of hospital admission was 39 days.
  • Most patients survived ECMO treatment. By the end of the study in September 2009, 21% of patients who received ECMO had died, 3% were still receiving ECMO, 6% were still in ICU but no longer receiving ECMO, 24% had been discharged from ICU but remained in hospital, and 47% had been discharged from hospital.
  • Although some of the patients had not yet been discharged from hospital, deaths after being taken off ECMO or after discharge from ICU were uncommon, so the death rate was not expected to change considerably.
  • Because the study was a case series that does not have a control group, we do not know how ECMO compares to other treatments. In most cases mechanical ventilation and other methods had been tried in these patients, so ECMO is likely to have been one of the few remaining treatment options.
  • The study did not assess long-term outcomes.

Where did the story come from?

This study was carried out by the Australia and New Zealand Extracorporeal Membrane Oxygenation (ANZ ECMO) Influenza Investigators. Sources of funding for the study were not reported. The study was published in the peer-reviewed Journal of the American Medical Association (JAMA ).

What kind of scientific study was this?

This case series analysed the outcomes of patients with severe swine flu-associated acute respiratory distress syndrome (ARDS) who had been treated with extracorporeal membrane oxygenation (ECMO). The study looked at patients in Intensive Care Units (ICUs) in Australia and New Zealand during their winter flu season.

ARDS is a severe complication that can occur in people with influenza. It occurs when the insides of the lungs become inflamed, leading to difficulties in getting oxygen into the blood stream and carbon dioxide waste out. The condition can be fatal and usually requires the patient to have mechanical ventilation (such as a ventilator) until their condition improves. ECMO is an alternative method that can be used if mechanical ventilation is not successful. It involves pumping the patient’s blood through a machine that oxygenates the blood and removes the carbon dioxide before pumping it back into the patient’s body.

The researchers contacted all 187 ICUs in Australia and New Zealand to identify all patients that had received ECMO for ARDS associated with strongly suspected or confirmed swine flu between June and August 2009. Children and adults were included, but not newborns. The search found 68 patients who had received ECMO in 15 intensive care units during this period. In addition, patients from the same ICUs who had confirmed influenza A but did not receive ECMO were identified. All patients were followed up to September 7 2009 to determine their outcomes. The researchers then calculated what proportion of people received ECMO, the characteristics of these patients, and their outcomes.

What were the results of the study?

When the researchers divided the number of people who received ECMO for severe flu-related ARDS (68 people) by the number of people in the general population, it represented 2.6 people per million who needed ECMO for confirmed or suspected swine flu.

Of these, about 78% (53 patients) were confirmed as having swine flu (influenza A (H1N1)); about 12%  were confirmed as having influenza A but the subtype not confirmed; and the remaining 10% had experienced symptoms of flu-like illness before developing ARDS and were suspected as having swine flu. In the same ICUs, 133 people with suspected or confirmed swine flu received mechanical ventilation, but not ECMO.

Patients receiving ECMO were aged 34.4 on average (median); three of the patients were children (younger than 15) and none were over 65. Almost all (97%) of the patients requiring ECMO fulfilled criteria for pneumonia. There was an even proportion of men and women. As well as suspected or confirmed swine flu, half of these patients were obese (BMI>30), 28% had asthma, and 15% had diabetes. Six patients (9%) were pregnant and  four patients (6%) had recently given birth. Over a quarter (28%) also had a secondary bacterial infection when they entered hospital.

The average (median) time between start of flu-like symptoms and ICU admission was five days, and it was an average of nine days between the onset of symptoms and ECMO. Oseltamivir (Tamiflu) had been used in 94% of the patients. The patients had severe respiratory failure that persisted even though they had received mechanical ventilation for an average of two days. Most patients (81%) had also received at least one other treatment for their ARDS before they started ECMO.

ECMO was given for an average (median) of 10 days (ranging from seven to 15 days).

During the study, 21% of the patients on ECMO died (14 out of 68 patients). By the end of the study, 9% of the patients (six patients) were still in the ICU, including two patients (3%) who were still receiving ECMO. Forty-eight (71%) of the patients had been successfully discharged from the ICU. Of these 48 patients, 32 had also been discharged from the hospital and 16 remained on a non-ICU ward in the hospital.

What interpretations did the researchers draw from these results?

The researchers concluded that one third of patients in ICUs with swine flu-related ARDS received ECMO in Australia and New Zealand in June to August 2009. These people were mainly young adults who had severely low oxygen levels in their blood. About a fifth of these patients died, and the authors report that these death rates are low considering how severe their condition was and the intensity of the treatment.

What does the NHS Knowledge Service make of this study?

This study has given an idea of how frequently ECMO might be required for cases of ARDS associated with swine flu. The study’s strengths include the fact that these patients are likely to represent all patients who received ECMO in Australia and New Zealand during the study period, so it is not likely to have been affected by biases in how patients were selected.

Such figures are useful for planning. Using these findings, the researchers estimate the European Union might need to provide ECMO to about 1,300 patients during the upcoming winter season.

The study does have some limitations, which the authors acknowledge:

  • The study collected data retrospectively, which can reduce the accuracy of findings. However, the researchers did try to counter this by using standardised data collection forms and definitions for various conditions, and trained research coordinators.
  • Only 78% of the patients had confirmed swine flu. The remainder were also very likely to have swine flu however, as they had either flu-like symptoms or confirmed influenza A at a time when swine flu was the main circulating flu strain.
  • As a case series, the study did not have a control group, and therefore it is not possible to know what would have happened to patients had they not received ECMO. However, the severity of their illness would have made it unethical not to provide what was considered to be the most appropriate treatment.
  • In order to publish the study in time for the northern hemisphere’s winter season, the study was finished in September, before all patients had been discharged from hospital. The outcomes of the patients remaining in hospital are therefore not known, and the death rate could actually be higher than that estimated. However, the authors note that most of the patients had been weaned from ECMO or discharged from ICU, and deaths after this stage in other patients were  uncommon. The study was also not able to look at the longer term outcomes of these patients, for example, how well their lungs continued to function.
  • The study does not offer an estimation of what proportion of people with swine flu will require ECMO, as it would have been difficult to determine exactly how many people had swine flu overall in the population during the period of the study.

NHS Attribution