- Email: [email protected]
Pneumonia cases following an EF-5 tornado
American Journal of Infection Control 43 (2015) 682-5
Contents lists available at ScienceDirect
American Journal of Infection Control
American Journal of Infection Control
journal homepage: www.ajicjournal.org
Major article
Pneumonia cases following an EF-5 tornado Beth A. Forshee-Hakala DO, PhD * Freeman Health System, Joplin, MO
Key Words: Natural disaster Respiratory infection
Background: Infections following a natural disaster such as an EF-5 tornado can be atypical and difficult to treat. Studies have looked at illness following several natural disasters, but few have studied respiratory illness following a tornado. Materials and Methods: A review of patients with pneumonia admitted during the period from May 22, 2009, through May 21, 2012, was completed. The Tornado Zone Group included adult patients who lived or worked in the tornado zone during the year following the tornado. Data were isolated by number of pneumonia cases within and outside the tornado zone per month per year. Results: An analysis of variance comparing the number of pneumonia cases from the tornado zone per month per year was significant at F2,38 ¼ 12.93 and P < .001, with increased cases in the Tornado Zone Group (P < .05). A t test comparing age of pneumonia patients found Tornado Zone patients to be younger than controls (t390 ¼ 5.14; P < .01). Microbes isolated from the Tornado Zone Group included uncommon pathogens not isolated during the 2 years prior. Discussion: The number of pneumonia cases may increase following tornadoes. Although current guidelines recommend narrow-spectrum antibiotics for community-acquired pneumonia, results of this study suggest the possible need for broader antimicrobial coverage after tornadoes. Copyright Ó 2015 by the Association for Professionals in Infection Control and Epidemiology, Inc. Published by Elsevier Inc. All rights reserved.
May 22, 2011, was undoubtedly 1 of the worst days in history for the town of Joplin, Missouri. It was sunny, hot, and humidda typical summer day. Then around 5:30 pm the sky darkened. Within minutes an EF-5 tornado with winds > 200 mph carved a path up to a mile wide and 6 miles long through town.1 With < 20 minutes’ warning, little preparation was possible and many were caught in its path. Thousands of structures were destroyed or severely damaged, and more than 150 people died. Remnants of buildings and vehicles were splattered with dirt and debris, a testament to the environmental changes that just took place. Freeman Hospital was immediately inundated with victims whose injures ranged from minor to grave. In fact, during the next 2 days, the hospital treated a number of patients that exceeded 2,000. For the next several weeks the entire town smelled of dust, mold, and destruction. Some victims left the area, whereas others worked to sift through what was left of their homes in search of any salvageable belongings. Volunteers came from all over the United
* Address correspondence to Beth A. Forshee-Hakala, DO, PhD, Meadville Medical Center, 751 Liberty St, Meadville, PA 16335. E-mail address: [email protected]. BF-H is currently affiliated with Meadville Medical Center, Meadville, PA; at the time the study was completed, she was with Freeman Health System, Joplin, MO. Conflicts of interest: None to report.
States to assist in the recovery efforts. Together, groups of people worked hours upon hours in the blazing sun cleaning up after this unbelievably destructive event. It is unknown exactly how many people worked to sift through the debris, but few wore masks or other protective equipment. This lack of protection undoubtedly caused inhalation of a multitude of respiratory irritants. Few studies have addressed infectious disease following tornadoes, although several studies describe disease following floodrelated disasters.2,3 From these studies, a classification system describing the pattern of infectious disease outbreak following natural disasters was created.4 Phase 1 (0-4 days) is referred to as the impact phase and includes the time when victims are extricated and traumatic injury is treated. Phase 2 (4 days to 4 weeks) is the time period when most air-, food-, and waterborne infections occur. Phase 3 (> 4 weeks) is referred to as the recovery phase, characterized by the onset of latent illnesses or infections with long incubation periods. It has been noted that the most common illnesses seen in phase 3 are those already endemic to the area.4 This result is to be expected given that disasters typically do not import new microbes but are capable of aerosolizing those already present in the environment. Some studies suggested an increase in respiratory illness among victims of natural disasters from mechanisms other than water ingestion.5-7 Many of these studies focus on the overcrowding
0196-6553/$36.00 - Copyright Ó 2015 by the Association for Professionals in Infection Control and Epidemiology, Inc. Published by Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.ajic.2015.02.027
B.A. Forshee-Hakala / American Journal of Infection Control 43 (2015) 682-5
conditions in relief shelters as the cause of illness, citing improvements in postdisaster living conditions as a necessity in reducing respiratory disease. Robinson et al8 looked closely at respiratory sequelae following various forms of natural disaster and described severe illness related to inhalation, aspiration, near drowning, and trauma. They reported that these illnesses were primarily caused by bronchospasm and/or respiratory failure. In that report, acute respiratory infections were not reported in those with inhalation injury but were limited to overcrowding and poor sanitation among victims. In contrast, Benedict and Park9 described respiratory fungal infections following a dust storm and an earthquake in California. Of the infections that resulted in a respiratory illness substantial enough to seek medical care, Coccidioides species was the primary microbe identified. During a tornado the powerful winds disrupt the environment by destroying structures and unearthing trees and soil. It would seem that such a powerful event would also aerosolize resident microbes of the soil. Although these microbes may not typically cause acute respiratory infection, it is plausible that the level of exposure following a tornado could lead to a significant increase in atypical respiratory infections. Therefore, the aim of this study was to compare pneumonia cases for patients who lived or worked within the tornado zone with those who did not during the year following the tornado. Further, these groups were also compared with those in the same geographic locations during the prior 2 years. Specifically, patient characteristics, differences in number of admissions, and isolated microbes were assessed. METHODS All aspects of this study were approved by the institutional review board at Freeman Health System, Joplin, Missouri. Patients Patients who were older than age 18 years and not pregnant at the time of admission were included in this study. Hospital charts were retrieved for all patients with an admitting diagnosis that included pneumonia during the period of May 22, 2009, through May 21, 2012. Cases were grouped into the following time periods: May 22, 2009, to May 21, 2010 (year 1); May 22, 2010, to May 21, 2011 (year 2); and May 22, 2011, to May 21, 2012 (year 3). For each year, pneumonia cases were divided into 2 groups: patients who lived or worked in the tornado zone and patients who neither lived nor worked in the tornado zone. The Tornado Zone Group consisted of the patients who lived or worked in the tornado zone during year 3. All other groups were considered Non-Tornado Zone Groups. A total of 1,296 charts were reviewed, of which 941 cases were determined to be true pneumonia and thus included in this study. Methods Charts were reviewed individually and data were recorded for age, gender, chest radiograph findings, procalcitonin level, white blood cell count, presence of cough and/or shortness of breath in the history and physical, blood/sputum culture findings, prior diagnosis of chronic obstructive pulmonary disease or asthma, and prior home oxygen use. Pneumonia cases were included if they had either radiologic evidence of infiltrate plus cough or elevated procalcitonin plus leukocytosis plus cough and shortness of breath at presentation. Sputum culture results were recorded when available only for patients from the tornado zone who met the pneumonia criteria. Data for number of cases per year per group were further divided by month of admission. In an effort to control for inflation
683
Table 1 Comparison of total admissions and pneumonia cases for patients within and outside the tornado zone during each of the 3 years studied
Year 1 Year 2 Year 3
Total admissions
Total cases
Tornado zone
Nontornado zone
13,972 14,168 16,015
314 235 392
132 (42) 92 (39) 191 (49)
182 (58) 143 (61) 201 (51)
NOTE. Data within parenthesis indicate percentage of total pneumonia cases for the corresponding year.
in total admits secondary to same-day surgeries, only admissions lasting longer than 24 hours were included in this study. Analysis Data for number of total admissions and total pneumonia cases were converted into percentage difference when comparing all 3 years. Data for pneumonia cases were further broken down into percentage of patients from within and outside the tornado zone. A 1-way analysis of variance was performed comparing the number of pneumonia cases per month for patients from within the tornado zone during each of the 3 years studied. Further analysis was performed using Scheffe post-hoc comparisons. An Independent t test was performed comparing the number of pneumonia cases per month for patients from within versus outside the tornado zone during year 3. Likewise, an Independent t test was performed comparing ages of pneumonia patients from within versus outside the tornado zone during years 1 and 2. Although sputum cultures were not available for all patients, specific microbes isolated were noted for patients from the tornado zone during all 3 years. RESULTS Total admissions to Freeman Hospital rose by 1.3% from year 1 to year 2, whereas there was a striking increase in total admissions of 11.5% from year 2 to year 3 (Table 1). Total pneumonia cases decreased by 25% from year 1 to year 2, but increased by 40.1% from year 2 to year 3. Further analysis comparing pneumonia cases between Years 1 and 3 revealed an increase of 20% in Year 3. Analysis of year 1 to year 2 revealed a 30% decrease in pneumonia cases among tornado zone patients and a 21% decrease among patients from outside the tornado zone. In contrast, a comparison of pneumonia cases for year 2 to year 3 revealed an increase of 51.8% in cases from the tornado zone and a 28.9% increase in cases from outside the tornado zone. A comparison of pneumonia cases between years 1 and 3 found an increase of 30.9% in pneumonia cases from within the tornado zone during year 3, whereas pneumonia cases increased only 9.5% from outside the tornado zone during this year. When looking closer at the pneumonia admissions from within the tornado zone, a 1-way analysis of variance comparing number of admissions per month for years 1 through 3 achieved significance at F2,38 ¼ 12.926 (P < .0001). Subsequent Scheffe post-hoc comparisons revealed that admissions from the tornado zone during year 3 were significantly increased when compared with each of years 1 and 2 (P < .05) (Fig 1). Similar comparisons between years 1 and 2 failed to reach significance. An Independent t test comparing number of pneumonia cases from the Tornado Zone Group versus the Non-Tornado Zone Group during year 3 failed to reach statistical significance. However, there was a nearly 40% increase in pneumonia cases in the Tornado Zone Group during the month of June, which was the month following the tornado (Fig 2). This number appeared to return to baseline during the month of July.
684
B.A. Forshee-Hakala / American Journal of Infection Control 43 (2015) 682-5 Table 2 Characteristics of patients admitted for pneumonia who lived or worked in the tornado zone during each of the 3 years studied
Year 1 Year 2 Year 3
Male
Female
COPD
Asthma
O2 dependent
67 (51) 44 (48) 107 (56)
65 (49) 48 (52) 84 (44)
38 (29) 21 (23) 42 (22)
1 (1) 0 4 (2)
5 (4) 6 (7) 8 (4)
NOTE. Data in parenthesis represent percentage of total pneumonia admissions for the corresponding year. COPD, chronic obstructive pulmonary disease; O2, oxygen.
Fig 1. Comparison of the number of pneumonia cases per month from patients who lived or worked in the tornado zone during years 1 through 3. Year 1 includes cases admitted between May 22, 2009, and May 21, 2010. Year 2 includes cases from May 22, 2010, to May 21, 2011, and year 3 represents cases from May 22, 2011, to May 21, 2012. Number of pneumonia cases in year 3 from the tornado zone was significantly greater than during years 1 and 2 (P < .05). Comparison of number of pneumonia cases in years 2 and 3 failed to reach significance (P > .05).
Table 3 Comparison of the microbes isolated in the sputum of patients from the tornado zone during years 1 and 2 versus year 3. Year 1 includes patients admitted between May 22, 2009, and May 21, 2010. Year 2 includes cases from May 22, 2010, to May 21, 2011, and year 3 represents cases from May 22, 2010, to May 21, 2012 Years 1 and 2 Streptococcus pneumoniae Staphylococcus aureus (including MRSA) Proteus mirabilis Streptococcus milleri group Moraxella catarrhalis Klebsiella pneumoniae Pseudomonas aeruginosa Haemophilus influenzae Mycoplasma pneumoniae
Year 3 Aspergillus* Pseudomonas putida* Enterococcus faecalis Enterobacter aerogenes (nonhospital acquired)* Alcaligenes xylosoxidans* Chlamydia pneumoniae Burkholderia species* Plus all microbes from years 1 and 2
MRSA, methicillin-resistant Staphylococcus aureus. *Microbe commonly found in soil and/or water.
DISCUSSION Fig 2. Comparison of the number of pneumonia cases from within the tornado zone versus outside the tornado zone during year 3 failed to reach significance (P > .05). Year 3 includes cases admitted between May 22, 2011, and May 21, 2012.
Table 2 describes patient characteristics for pneumonia cases from the tornado zone over all 3 years studied. Whereas the number of male versus female patients from within the tornado zone during years 1 and 2 were very similar, a greater proportion of males (56%) versus females (44%) from the tornado zone were admitted during year 3. The percentage of patients within the tornado zone who had a previous diagnosis of chronic obstructive pulmonary disease or asthma were similar across the 3 years examined, with years 2 and 3 being the lowest. Further, the percentage of patients who were oxygen-dependent before admission was similar across the 3 years for patients from the tornado zone. An Independent t test comparing age of pneumonia cases from within versus outside the tornado zone during year 3 revealed that patients from the tornado zone were younger than those from outside this area (t390 ¼ 5.14; P < .01). Similar comparisons for the 2 years prior failed to reach significance. The average age for the Tornado Zone Group was 62 years compared with an average age range of 70-73.5 years in Non-Tornado Zone patients. The majority of pneumonia cases in which sputum was collected during the 3 years studied had negative cultures, but comparison of positive cultures revealed some atypical microbes in patients from the Tornado Zone Group compared wit Non-Tornado Zone patients during years 1 and 2 (Table 3). Some of these pathogens are known to be common in soil, vegetation and/or water, including Alcaligenes xylosoxidans, Burkholderia species, Aspergillus, Pseudomonas putida, and Enterobacter aerogenes.
The most striking finding of this study is that the number of pneumonia cases from within the tornado zone was drastically increased during the month following the tornado compared with prior years, and that these patients were younger than pneumonia patients from outside the tornado zone. Further, the microbes isolated from the sputum of patients from the Tornado Zone Group compared with those who lived or worked in the tornado zone 2 years prior included several atypical pathogens commonly found in soil, decaying vegetation, and water. Freeman Hospital is a 446-bed tertiary care center that together with the 367-bed St John’s Mercy Hospital, cared for an estimated 450,000 people in a 4-state region before the May 22, 2011, tornado. Therefore, an increase in admissions following the tornado was expected due to the destruction of St John’s Mercy Hospital. It could be that the increase in pneumonia cases was secondary to an increase in patient volume. However, the number of patients with pneumonia admitted from within the tornado zone during year 3 was 51.8% greater than year 2, with only a 28.9% increase in cases from outside the tornado zone. Further, people from outside communities came to the aid of Joplin following the tornado, making it uncertain whether some of the pneumonia cases in the Non-Tornado Zone group were secondary to exposure to the dust and debris during relief efforts. Further, many of the victims of the tornado left the Joplin area, and their health status is unknown. Taken together, this evidence suggests a strong link between exposure to the environmental disturbance produced by the tornado and an increase in pneumonia cases. In fact, it is highly likely that the data here underestimate the actual pneumonia cases in the Tornado Zone Group because they do not take into account those who left the area or were treated elsewhere.
B.A. Forshee-Hakala / American Journal of Infection Control 43 (2015) 682-5
When looking at the number of pneumonia cases from within the tornado zone, there were more cases in the Tornado Zone Group during most months compared with groups from the same geographic location during years 1 and 2. In particular, the month of June in year 3 saw a large spike in pneumonia admissions compared with years 1 and 2. This pattern was described in the articles by Kouadio et al4 and Aghababian and Teuscher5 as phase 2 in the trend of infectious disease following natural disasters. Phase 2 is that time period between 4 days and 4 weeks following a disaster, when most air-, food-, and waterborne illnesses present. However, their studies proposed that these illnesses were primarily related to overcrowded relief shelters rather than the disaster itself. Following the Joplin, Missouri, EF-5 tornado, unclean and crowded relief shelters were not a problem. Many victims were quickly provided with or obtained private housing. Therefore, results presented here suggest that the increase in pneumonia cases was a direct result of the disaster itself. Moreover, the microbes isolated from the sputum of these victims compared with pneumonia patients from the same geographic location during the prior 2 years included several that are frequently found in the soil, water, or vegetation. This finding could provide further support that these pneumonia cases are the direct result of the aerosolization of microbes during the tornado. Before conducting the study, we postulated that the age of pneumonia patient in the Tornado Zone Group would be lower than the 2 years prior. The Tornado Zone group age averaged 62 years, whereas the Non-Tornado Zone group ages ranged from a mean of 70-73.5 years. This finding supports anecdotal reports describing more serious pneumonia cases in younger patients following the tornado. Combined with the atypical organisms isolated in the sputum of the Tornado Zone Group, these findings support the notion that the pneumonia cases following the tornado were more severe than those encountered during the prior 2 years. Further, the resistance patterns of these microbes made treatment difficult and frequently required escalation of antimicrobial agents to cover a broad spectrum of pathogens. Current guidelines of the Infectious Diseases Society of America recommend several regimens of empiric antibiotics for cases of community-acquired pneumonia depending on comorbidities and prior antibiotic use.10 However, these recommendations are most often narrow in spectrum. Although this is appropriate for nearly all community-acquired pneumonia cases, pneumonia cases occurring in victims of natural disasters causing soil disruption should be treated differently regardless of age and comorbidities. Results of this study show that a more aggressive approach to treating pneumonia in tornado victims should be taken, including broader coverage for gram-negative and atypical microbes. Therefore, it is recommended that empiric antimicrobial coverage in these patients include multidrug coverage against gram-positive, including methicillin-resistant Staphylococcus aureus; gram-negative; and atypical microbes. Further, sputum cultures should be collected on all patients with pneumonia, and antibiotics should then be quickly de-escalated once an organism is identified. This approach is frequently frowned upon given an increase in antimicrobial resistance patterns. However, it seems that this approach is warranted in special cases such as postdisaster pneumonias where polymicrobial infections with atypical organisms are expected. Additionally, the local health department should embark on a program of close monitoring for outbreaks of respiratory illnesses with atypical microbes after tornadoes and other natural disasters. Noji11 described the importance of epidemiologic information to the planning for and response following a variety of natural
685
disasters. Several studies were described that looked at patterns of injury and illness following tornadoes. Fractures, lacerations, penetrating trauma, and subsequent development of wound infections and sepsis were most common. This study suggests that surveillance of respiratory illness following tornadoes should be included in public health reporting and monitoring, and that such information could be useful in planning and implementing care. Limitations of this study are that it is a retrospective case study in which chart review was the mechanism of data collection. To overcome this weakness, year-to-year fluctuation in pneumonia admissions were controlled for by examining cases from the 2 years before the tornado for comparison. This 3-year time frame reduces the chance of error related to natural fluctuation. Other limitations are that the data represented fewer than 400 cases per year due to the size of the hospital. It is likely that many patients who developed pneumonia secondary to the tornado went elsewhere for care as a result of the long wait lines for the emergency department and floor beds. However, this could also make the data here stronger because additional pneumonia cases would most likely strengthen the findings. CONCLUSIONS This study is an important addition to literature describing natural disaster-related illness. Data presented here suggest that the disaster itself, especially in the case of tornadoes, can disturb the environment in such a way that pneumonia can ensue. This finding is important to those who are caring for victims of tornadoes, in that they must be cognizant of the atypical microbes present in the environment that can cause illness. Further, decisions for antimicrobial coverage in these patients should be made with broad coverage in mind, regardless of patient characteristics. Hospitals and local health departments should have a protocol for identifying and reporting respiratory infections from atypical microbes following tornadoes. This practice could provide valuable guidance for the selection of effective antimicrobial treatment. References 1. U.S. Department of Commerce. NWS Central Region Service Assessment. Joplin, Missouri, Tornado-May 22, 2011. Kansas City, Missouri: NWS Central Region Headquarters; July 2011. 2. Centers for Disease Control and Prevention (CDC). Infectious Disease and Dermatologic Conditions in Evacuees and Rescue Workers after Hurricane Katrina e Multiple states, August e September, 2005. MMWR Morb Mortal Wkly Rep 2005;54:961-4. 3. Uckay I, Sax H, Harbarth S, Bernard L, Pittet D. Multi-resistant infections in Repatriated patients after natural disasters: Lessons Learned from the 2004 Tsunami for hospital infection Non-tornado zone. J Hosp Infect 2008;68:1-8. 4. Kouadio IK, Alijunid S, Kamigaki T, Hammad K, Oshitani H. Infectious diseases following natural disasters: Prevention and Non-tornado zone Measures. 2012. Expert Rev Anti Infect Ther 2012;10:95-104. 5. Aghababian RV, Teuscher J. Infectious diseases following major disasters. Ann Emerg Med 1992;21:4. 6. Steinberg SM, Nichols RL. Infections and sepsis in disasters. Disaster Management 1991;7:437-50. 7. Ivy J. Infections Encountered in tornado and Automobile Accident victims. J Indiana State Med Assoc 1968;61:1657. 8. Robinson B, Alatas MF, Robertson A, Steer H. Natural disasters and the Lung. Respirology 2011;16:386-95. 9. Benedict K, Park BJ. Invasive fungal infections after natural disasters. Emerg Infect Dis 2014;20:349-55. 10. Mandell LA, Wunderink RG, Anzueto A, Bartlett JG, Campbell GD, Dean NC, et al, Infectious Diseases Society of America; American Thoracic Society. Consensus guidelines on the Management of community-acquired pneumonia in adults. Clin Infect Dis 2007;44:S27-72. 11. Noji EK. The public health Consequences of disasters. Public Health and Disasters 2000;15:147-57.
Contents lists available at ScienceDirect
American Journal of Infection Control
American Journal of Infection Control
journal homepage: www.ajicjournal.org
Major article
Pneumonia cases following an EF-5 tornado Beth A. Forshee-Hakala DO, PhD * Freeman Health System, Joplin, MO
Key Words: Natural disaster Respiratory infection
Background: Infections following a natural disaster such as an EF-5 tornado can be atypical and difficult to treat. Studies have looked at illness following several natural disasters, but few have studied respiratory illness following a tornado. Materials and Methods: A review of patients with pneumonia admitted during the period from May 22, 2009, through May 21, 2012, was completed. The Tornado Zone Group included adult patients who lived or worked in the tornado zone during the year following the tornado. Data were isolated by number of pneumonia cases within and outside the tornado zone per month per year. Results: An analysis of variance comparing the number of pneumonia cases from the tornado zone per month per year was significant at F2,38 ¼ 12.93 and P < .001, with increased cases in the Tornado Zone Group (P < .05). A t test comparing age of pneumonia patients found Tornado Zone patients to be younger than controls (t390 ¼ 5.14; P < .01). Microbes isolated from the Tornado Zone Group included uncommon pathogens not isolated during the 2 years prior. Discussion: The number of pneumonia cases may increase following tornadoes. Although current guidelines recommend narrow-spectrum antibiotics for community-acquired pneumonia, results of this study suggest the possible need for broader antimicrobial coverage after tornadoes. Copyright Ó 2015 by the Association for Professionals in Infection Control and Epidemiology, Inc. Published by Elsevier Inc. All rights reserved.
May 22, 2011, was undoubtedly 1 of the worst days in history for the town of Joplin, Missouri. It was sunny, hot, and humidda typical summer day. Then around 5:30 pm the sky darkened. Within minutes an EF-5 tornado with winds > 200 mph carved a path up to a mile wide and 6 miles long through town.1 With < 20 minutes’ warning, little preparation was possible and many were caught in its path. Thousands of structures were destroyed or severely damaged, and more than 150 people died. Remnants of buildings and vehicles were splattered with dirt and debris, a testament to the environmental changes that just took place. Freeman Hospital was immediately inundated with victims whose injures ranged from minor to grave. In fact, during the next 2 days, the hospital treated a number of patients that exceeded 2,000. For the next several weeks the entire town smelled of dust, mold, and destruction. Some victims left the area, whereas others worked to sift through what was left of their homes in search of any salvageable belongings. Volunteers came from all over the United
* Address correspondence to Beth A. Forshee-Hakala, DO, PhD, Meadville Medical Center, 751 Liberty St, Meadville, PA 16335. E-mail address: [email protected]. BF-H is currently affiliated with Meadville Medical Center, Meadville, PA; at the time the study was completed, she was with Freeman Health System, Joplin, MO. Conflicts of interest: None to report.
States to assist in the recovery efforts. Together, groups of people worked hours upon hours in the blazing sun cleaning up after this unbelievably destructive event. It is unknown exactly how many people worked to sift through the debris, but few wore masks or other protective equipment. This lack of protection undoubtedly caused inhalation of a multitude of respiratory irritants. Few studies have addressed infectious disease following tornadoes, although several studies describe disease following floodrelated disasters.2,3 From these studies, a classification system describing the pattern of infectious disease outbreak following natural disasters was created.4 Phase 1 (0-4 days) is referred to as the impact phase and includes the time when victims are extricated and traumatic injury is treated. Phase 2 (4 days to 4 weeks) is the time period when most air-, food-, and waterborne infections occur. Phase 3 (> 4 weeks) is referred to as the recovery phase, characterized by the onset of latent illnesses or infections with long incubation periods. It has been noted that the most common illnesses seen in phase 3 are those already endemic to the area.4 This result is to be expected given that disasters typically do not import new microbes but are capable of aerosolizing those already present in the environment. Some studies suggested an increase in respiratory illness among victims of natural disasters from mechanisms other than water ingestion.5-7 Many of these studies focus on the overcrowding
0196-6553/$36.00 - Copyright Ó 2015 by the Association for Professionals in Infection Control and Epidemiology, Inc. Published by Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.ajic.2015.02.027
B.A. Forshee-Hakala / American Journal of Infection Control 43 (2015) 682-5
conditions in relief shelters as the cause of illness, citing improvements in postdisaster living conditions as a necessity in reducing respiratory disease. Robinson et al8 looked closely at respiratory sequelae following various forms of natural disaster and described severe illness related to inhalation, aspiration, near drowning, and trauma. They reported that these illnesses were primarily caused by bronchospasm and/or respiratory failure. In that report, acute respiratory infections were not reported in those with inhalation injury but were limited to overcrowding and poor sanitation among victims. In contrast, Benedict and Park9 described respiratory fungal infections following a dust storm and an earthquake in California. Of the infections that resulted in a respiratory illness substantial enough to seek medical care, Coccidioides species was the primary microbe identified. During a tornado the powerful winds disrupt the environment by destroying structures and unearthing trees and soil. It would seem that such a powerful event would also aerosolize resident microbes of the soil. Although these microbes may not typically cause acute respiratory infection, it is plausible that the level of exposure following a tornado could lead to a significant increase in atypical respiratory infections. Therefore, the aim of this study was to compare pneumonia cases for patients who lived or worked within the tornado zone with those who did not during the year following the tornado. Further, these groups were also compared with those in the same geographic locations during the prior 2 years. Specifically, patient characteristics, differences in number of admissions, and isolated microbes were assessed. METHODS All aspects of this study were approved by the institutional review board at Freeman Health System, Joplin, Missouri. Patients Patients who were older than age 18 years and not pregnant at the time of admission were included in this study. Hospital charts were retrieved for all patients with an admitting diagnosis that included pneumonia during the period of May 22, 2009, through May 21, 2012. Cases were grouped into the following time periods: May 22, 2009, to May 21, 2010 (year 1); May 22, 2010, to May 21, 2011 (year 2); and May 22, 2011, to May 21, 2012 (year 3). For each year, pneumonia cases were divided into 2 groups: patients who lived or worked in the tornado zone and patients who neither lived nor worked in the tornado zone. The Tornado Zone Group consisted of the patients who lived or worked in the tornado zone during year 3. All other groups were considered Non-Tornado Zone Groups. A total of 1,296 charts were reviewed, of which 941 cases were determined to be true pneumonia and thus included in this study. Methods Charts were reviewed individually and data were recorded for age, gender, chest radiograph findings, procalcitonin level, white blood cell count, presence of cough and/or shortness of breath in the history and physical, blood/sputum culture findings, prior diagnosis of chronic obstructive pulmonary disease or asthma, and prior home oxygen use. Pneumonia cases were included if they had either radiologic evidence of infiltrate plus cough or elevated procalcitonin plus leukocytosis plus cough and shortness of breath at presentation. Sputum culture results were recorded when available only for patients from the tornado zone who met the pneumonia criteria. Data for number of cases per year per group were further divided by month of admission. In an effort to control for inflation
683
Table 1 Comparison of total admissions and pneumonia cases for patients within and outside the tornado zone during each of the 3 years studied
Year 1 Year 2 Year 3
Total admissions
Total cases
Tornado zone
Nontornado zone
13,972 14,168 16,015
314 235 392
132 (42) 92 (39) 191 (49)
182 (58) 143 (61) 201 (51)
NOTE. Data within parenthesis indicate percentage of total pneumonia cases for the corresponding year.
in total admits secondary to same-day surgeries, only admissions lasting longer than 24 hours were included in this study. Analysis Data for number of total admissions and total pneumonia cases were converted into percentage difference when comparing all 3 years. Data for pneumonia cases were further broken down into percentage of patients from within and outside the tornado zone. A 1-way analysis of variance was performed comparing the number of pneumonia cases per month for patients from within the tornado zone during each of the 3 years studied. Further analysis was performed using Scheffe post-hoc comparisons. An Independent t test was performed comparing the number of pneumonia cases per month for patients from within versus outside the tornado zone during year 3. Likewise, an Independent t test was performed comparing ages of pneumonia patients from within versus outside the tornado zone during years 1 and 2. Although sputum cultures were not available for all patients, specific microbes isolated were noted for patients from the tornado zone during all 3 years. RESULTS Total admissions to Freeman Hospital rose by 1.3% from year 1 to year 2, whereas there was a striking increase in total admissions of 11.5% from year 2 to year 3 (Table 1). Total pneumonia cases decreased by 25% from year 1 to year 2, but increased by 40.1% from year 2 to year 3. Further analysis comparing pneumonia cases between Years 1 and 3 revealed an increase of 20% in Year 3. Analysis of year 1 to year 2 revealed a 30% decrease in pneumonia cases among tornado zone patients and a 21% decrease among patients from outside the tornado zone. In contrast, a comparison of pneumonia cases for year 2 to year 3 revealed an increase of 51.8% in cases from the tornado zone and a 28.9% increase in cases from outside the tornado zone. A comparison of pneumonia cases between years 1 and 3 found an increase of 30.9% in pneumonia cases from within the tornado zone during year 3, whereas pneumonia cases increased only 9.5% from outside the tornado zone during this year. When looking closer at the pneumonia admissions from within the tornado zone, a 1-way analysis of variance comparing number of admissions per month for years 1 through 3 achieved significance at F2,38 ¼ 12.926 (P < .0001). Subsequent Scheffe post-hoc comparisons revealed that admissions from the tornado zone during year 3 were significantly increased when compared with each of years 1 and 2 (P < .05) (Fig 1). Similar comparisons between years 1 and 2 failed to reach significance. An Independent t test comparing number of pneumonia cases from the Tornado Zone Group versus the Non-Tornado Zone Group during year 3 failed to reach statistical significance. However, there was a nearly 40% increase in pneumonia cases in the Tornado Zone Group during the month of June, which was the month following the tornado (Fig 2). This number appeared to return to baseline during the month of July.
684
B.A. Forshee-Hakala / American Journal of Infection Control 43 (2015) 682-5 Table 2 Characteristics of patients admitted for pneumonia who lived or worked in the tornado zone during each of the 3 years studied
Year 1 Year 2 Year 3
Male
Female
COPD
Asthma
O2 dependent
67 (51) 44 (48) 107 (56)
65 (49) 48 (52) 84 (44)
38 (29) 21 (23) 42 (22)
1 (1) 0 4 (2)
5 (4) 6 (7) 8 (4)
NOTE. Data in parenthesis represent percentage of total pneumonia admissions for the corresponding year. COPD, chronic obstructive pulmonary disease; O2, oxygen.
Fig 1. Comparison of the number of pneumonia cases per month from patients who lived or worked in the tornado zone during years 1 through 3. Year 1 includes cases admitted between May 22, 2009, and May 21, 2010. Year 2 includes cases from May 22, 2010, to May 21, 2011, and year 3 represents cases from May 22, 2011, to May 21, 2012. Number of pneumonia cases in year 3 from the tornado zone was significantly greater than during years 1 and 2 (P < .05). Comparison of number of pneumonia cases in years 2 and 3 failed to reach significance (P > .05).
Table 3 Comparison of the microbes isolated in the sputum of patients from the tornado zone during years 1 and 2 versus year 3. Year 1 includes patients admitted between May 22, 2009, and May 21, 2010. Year 2 includes cases from May 22, 2010, to May 21, 2011, and year 3 represents cases from May 22, 2010, to May 21, 2012 Years 1 and 2 Streptococcus pneumoniae Staphylococcus aureus (including MRSA) Proteus mirabilis Streptococcus milleri group Moraxella catarrhalis Klebsiella pneumoniae Pseudomonas aeruginosa Haemophilus influenzae Mycoplasma pneumoniae
Year 3 Aspergillus* Pseudomonas putida* Enterococcus faecalis Enterobacter aerogenes (nonhospital acquired)* Alcaligenes xylosoxidans* Chlamydia pneumoniae Burkholderia species* Plus all microbes from years 1 and 2
MRSA, methicillin-resistant Staphylococcus aureus. *Microbe commonly found in soil and/or water.
DISCUSSION Fig 2. Comparison of the number of pneumonia cases from within the tornado zone versus outside the tornado zone during year 3 failed to reach significance (P > .05). Year 3 includes cases admitted between May 22, 2011, and May 21, 2012.
Table 2 describes patient characteristics for pneumonia cases from the tornado zone over all 3 years studied. Whereas the number of male versus female patients from within the tornado zone during years 1 and 2 were very similar, a greater proportion of males (56%) versus females (44%) from the tornado zone were admitted during year 3. The percentage of patients within the tornado zone who had a previous diagnosis of chronic obstructive pulmonary disease or asthma were similar across the 3 years examined, with years 2 and 3 being the lowest. Further, the percentage of patients who were oxygen-dependent before admission was similar across the 3 years for patients from the tornado zone. An Independent t test comparing age of pneumonia cases from within versus outside the tornado zone during year 3 revealed that patients from the tornado zone were younger than those from outside this area (t390 ¼ 5.14; P < .01). Similar comparisons for the 2 years prior failed to reach significance. The average age for the Tornado Zone Group was 62 years compared with an average age range of 70-73.5 years in Non-Tornado Zone patients. The majority of pneumonia cases in which sputum was collected during the 3 years studied had negative cultures, but comparison of positive cultures revealed some atypical microbes in patients from the Tornado Zone Group compared wit Non-Tornado Zone patients during years 1 and 2 (Table 3). Some of these pathogens are known to be common in soil, vegetation and/or water, including Alcaligenes xylosoxidans, Burkholderia species, Aspergillus, Pseudomonas putida, and Enterobacter aerogenes.
The most striking finding of this study is that the number of pneumonia cases from within the tornado zone was drastically increased during the month following the tornado compared with prior years, and that these patients were younger than pneumonia patients from outside the tornado zone. Further, the microbes isolated from the sputum of patients from the Tornado Zone Group compared with those who lived or worked in the tornado zone 2 years prior included several atypical pathogens commonly found in soil, decaying vegetation, and water. Freeman Hospital is a 446-bed tertiary care center that together with the 367-bed St John’s Mercy Hospital, cared for an estimated 450,000 people in a 4-state region before the May 22, 2011, tornado. Therefore, an increase in admissions following the tornado was expected due to the destruction of St John’s Mercy Hospital. It could be that the increase in pneumonia cases was secondary to an increase in patient volume. However, the number of patients with pneumonia admitted from within the tornado zone during year 3 was 51.8% greater than year 2, with only a 28.9% increase in cases from outside the tornado zone. Further, people from outside communities came to the aid of Joplin following the tornado, making it uncertain whether some of the pneumonia cases in the Non-Tornado Zone group were secondary to exposure to the dust and debris during relief efforts. Further, many of the victims of the tornado left the Joplin area, and their health status is unknown. Taken together, this evidence suggests a strong link between exposure to the environmental disturbance produced by the tornado and an increase in pneumonia cases. In fact, it is highly likely that the data here underestimate the actual pneumonia cases in the Tornado Zone Group because they do not take into account those who left the area or were treated elsewhere.
B.A. Forshee-Hakala / American Journal of Infection Control 43 (2015) 682-5
When looking at the number of pneumonia cases from within the tornado zone, there were more cases in the Tornado Zone Group during most months compared with groups from the same geographic location during years 1 and 2. In particular, the month of June in year 3 saw a large spike in pneumonia admissions compared with years 1 and 2. This pattern was described in the articles by Kouadio et al4 and Aghababian and Teuscher5 as phase 2 in the trend of infectious disease following natural disasters. Phase 2 is that time period between 4 days and 4 weeks following a disaster, when most air-, food-, and waterborne illnesses present. However, their studies proposed that these illnesses were primarily related to overcrowded relief shelters rather than the disaster itself. Following the Joplin, Missouri, EF-5 tornado, unclean and crowded relief shelters were not a problem. Many victims were quickly provided with or obtained private housing. Therefore, results presented here suggest that the increase in pneumonia cases was a direct result of the disaster itself. Moreover, the microbes isolated from the sputum of these victims compared with pneumonia patients from the same geographic location during the prior 2 years included several that are frequently found in the soil, water, or vegetation. This finding could provide further support that these pneumonia cases are the direct result of the aerosolization of microbes during the tornado. Before conducting the study, we postulated that the age of pneumonia patient in the Tornado Zone Group would be lower than the 2 years prior. The Tornado Zone group age averaged 62 years, whereas the Non-Tornado Zone group ages ranged from a mean of 70-73.5 years. This finding supports anecdotal reports describing more serious pneumonia cases in younger patients following the tornado. Combined with the atypical organisms isolated in the sputum of the Tornado Zone Group, these findings support the notion that the pneumonia cases following the tornado were more severe than those encountered during the prior 2 years. Further, the resistance patterns of these microbes made treatment difficult and frequently required escalation of antimicrobial agents to cover a broad spectrum of pathogens. Current guidelines of the Infectious Diseases Society of America recommend several regimens of empiric antibiotics for cases of community-acquired pneumonia depending on comorbidities and prior antibiotic use.10 However, these recommendations are most often narrow in spectrum. Although this is appropriate for nearly all community-acquired pneumonia cases, pneumonia cases occurring in victims of natural disasters causing soil disruption should be treated differently regardless of age and comorbidities. Results of this study show that a more aggressive approach to treating pneumonia in tornado victims should be taken, including broader coverage for gram-negative and atypical microbes. Therefore, it is recommended that empiric antimicrobial coverage in these patients include multidrug coverage against gram-positive, including methicillin-resistant Staphylococcus aureus; gram-negative; and atypical microbes. Further, sputum cultures should be collected on all patients with pneumonia, and antibiotics should then be quickly de-escalated once an organism is identified. This approach is frequently frowned upon given an increase in antimicrobial resistance patterns. However, it seems that this approach is warranted in special cases such as postdisaster pneumonias where polymicrobial infections with atypical organisms are expected. Additionally, the local health department should embark on a program of close monitoring for outbreaks of respiratory illnesses with atypical microbes after tornadoes and other natural disasters. Noji11 described the importance of epidemiologic information to the planning for and response following a variety of natural
685
disasters. Several studies were described that looked at patterns of injury and illness following tornadoes. Fractures, lacerations, penetrating trauma, and subsequent development of wound infections and sepsis were most common. This study suggests that surveillance of respiratory illness following tornadoes should be included in public health reporting and monitoring, and that such information could be useful in planning and implementing care. Limitations of this study are that it is a retrospective case study in which chart review was the mechanism of data collection. To overcome this weakness, year-to-year fluctuation in pneumonia admissions were controlled for by examining cases from the 2 years before the tornado for comparison. This 3-year time frame reduces the chance of error related to natural fluctuation. Other limitations are that the data represented fewer than 400 cases per year due to the size of the hospital. It is likely that many patients who developed pneumonia secondary to the tornado went elsewhere for care as a result of the long wait lines for the emergency department and floor beds. However, this could also make the data here stronger because additional pneumonia cases would most likely strengthen the findings. CONCLUSIONS This study is an important addition to literature describing natural disaster-related illness. Data presented here suggest that the disaster itself, especially in the case of tornadoes, can disturb the environment in such a way that pneumonia can ensue. This finding is important to those who are caring for victims of tornadoes, in that they must be cognizant of the atypical microbes present in the environment that can cause illness. Further, decisions for antimicrobial coverage in these patients should be made with broad coverage in mind, regardless of patient characteristics. Hospitals and local health departments should have a protocol for identifying and reporting respiratory infections from atypical microbes following tornadoes. This practice could provide valuable guidance for the selection of effective antimicrobial treatment. References 1. U.S. Department of Commerce. NWS Central Region Service Assessment. Joplin, Missouri, Tornado-May 22, 2011. Kansas City, Missouri: NWS Central Region Headquarters; July 2011. 2. Centers for Disease Control and Prevention (CDC). Infectious Disease and Dermatologic Conditions in Evacuees and Rescue Workers after Hurricane Katrina e Multiple states, August e September, 2005. MMWR Morb Mortal Wkly Rep 2005;54:961-4. 3. Uckay I, Sax H, Harbarth S, Bernard L, Pittet D. Multi-resistant infections in Repatriated patients after natural disasters: Lessons Learned from the 2004 Tsunami for hospital infection Non-tornado zone. J Hosp Infect 2008;68:1-8. 4. Kouadio IK, Alijunid S, Kamigaki T, Hammad K, Oshitani H. Infectious diseases following natural disasters: Prevention and Non-tornado zone Measures. 2012. Expert Rev Anti Infect Ther 2012;10:95-104. 5. Aghababian RV, Teuscher J. Infectious diseases following major disasters. Ann Emerg Med 1992;21:4. 6. Steinberg SM, Nichols RL. Infections and sepsis in disasters. Disaster Management 1991;7:437-50. 7. Ivy J. Infections Encountered in tornado and Automobile Accident victims. J Indiana State Med Assoc 1968;61:1657. 8. Robinson B, Alatas MF, Robertson A, Steer H. Natural disasters and the Lung. Respirology 2011;16:386-95. 9. Benedict K, Park BJ. Invasive fungal infections after natural disasters. Emerg Infect Dis 2014;20:349-55. 10. Mandell LA, Wunderink RG, Anzueto A, Bartlett JG, Campbell GD, Dean NC, et al, Infectious Diseases Society of America; American Thoracic Society. Consensus guidelines on the Management of community-acquired pneumonia in adults. Clin Infect Dis 2007;44:S27-72. 11. Noji EK. The public health Consequences of disasters. Public Health and Disasters 2000;15:147-57.