Survival From Severe Pandemic H1N1 in Urban and Rural Turkey: A Case Series

Discussion

To our knowledge, this is the first report of data on the characteristics of critically ill H1N1 patients in Turkey, in whom H1N1 predominantly affected young adults and women. The most common symptoms were cough, fever, myalgia, and headache. Critical illness occurred within the first week after the onset of symptoms. All the patients who needed intensive care had hypoxemia (P_aO2/F_IO2 < 300 mm Hg), and 90% of the intubated patients died. Our measures of illness severity (APACHE II, severity of hypoxemia, pH, and NIV failure), time to H1N1 confirmation, and rural dwelling were predictors of mortality.

Since the beginning of the outbreak there have been a few reports on H1N1 patients treated in the ICU. The mortality rates have ranged from 17% to 50%.^6,8,9,13,14 In our patients the overall mortality at day 28 was 50%, which may be due to the lack of ICU facilities in rural Turkey, because 78% of the nonsurvivors were from rural areas. Critical care medicine is still not a specialty in Turkey, and most of the hospitals in the rural areas do not have optimal ICUs. Pulmonary physicians, anesthesiologists, and internal medicine specialists are in charge of the ICUs, but there is no standard fellowship program. Most of the critically ill patients are referred to the ICUs in the urban hospitals, because the rural hospitals lack experienced ICU staff and technical facilities such as advanced ventilators and monitoring systems. All our rural patients in this study were transferred to our hospital from rural hospitals, so they did not receive ICU treatment until they were at our hospital, which may have delayed antiviral therapy and other procedures, thus increasing the mortality in the rural patients. Additionally, our rural patients had much lower income than our urban patients, which might have affected their nutritional status and thus their immune status. NIV success/failure delineated the more severely ill patients (who failed NIV) from those with milder disease (who tolerated NIV). H1N1 confirmation time was also longer in nonsurvivors than in survivors, and later diagnosis might have affected the access time to advanced ICU care and procedures.

The overall NIV success rate in our patient group was 30%. Rello and co-workers reported an NIV success rate of 25% in Spain.⁸ In other reports, from Canada and Utah, the NIV success rate was about 15%.⁹ NIV predominantly benefits hypercapnic respiratory failure, although some studies have shown good results in hypoxemic patients also.^15–17 The benefit of NIV was mostly in the subgroup of patients with severe community-acquired pneumonia: not in those with ARDS/ALI.¹⁶ A meta-analysis suggested that NIV did not decrease the intubation rate and that there was not enough evidence to support the use of NIV in ALI/ARDS patients.¹⁸ Recent research suggests that NIV benefits a small percentage of patients with ALI/ARDS due to H1N1, but the past high NIV failure rate should be kept in mind, and these patients should be closely monitored and intubated without delay if they clinically deteriorate.¹⁹

The H1N1 pandemic differed from earlier pandemics such as the H2N2 1889 Russian influenza pandemic, the H2N2 1957 Asian influenza pandemic, and the H3N2 1968 Hong Kong influenza pandemic, in which the infection was mostly among children and young adults, and the mortality was greatest in very young children and the elderly.²⁰ In the H1N1 pandemic, severe illness requiring intensive care was predominantly in adults, 20–70 years of age. and the median age of nonsurvivors was only 39 years. The 1918 Spanish influenza pandemic was similar and was also caused by the H1N1 subtype.²¹ Most fatalities in the 1918 Spanish pandemic had secondary bacterial pneumonia and ARDS. It is still unknown why so many deaths occurred among the young and healthy. In the current H1N1 pandemic, people over age 70 were not so severely affected, which might be due to previous exposure to H1N1 viruses that circulated between 1918 and 1957.²²

The most common underlying condition that correlated with severe H1N1 illness in our patients was obesity, with a rate of 33%. Similar findings came from Mexico, Spain, and Canada.^6–8 Primary risk factors were chronic lung disease, cardiac diseases, and pregnancy. However, besides those underlying conditions, 7 of our 18 patients were previously healthy adults, and two of them died.

The recent H1N1 pandemic seemed to be associated with a high incidence of severe disease and substantial mortality in pregnant women, whose hospital admission rate was 4 times higher than that of the general population.²³ Other reports found higher ICU admission due to H1N1 in pregnant women than in nonpregnant women.^24,25 Clinicians caring for pregnant women should be alert for influenza-like symptoms when a pandemic virus is locally present, and should encourage pregnant women to be vaccinated and to begin empirical antiviral therapy immediately if influenza symptoms occur, irrespective of negative diagnostic tests.

In several countries, pulmonary embolism has been present in H1N1 patients at admission or arose during ICU stay,^26–28 and two of our patients had pulmonary embolism confirmed via contrast-enhanced computed tomogram.