The Epidemiology of Acute Respiratory Infections

Kenrad E. Nelson

Mark C. Steinhoff

INTRODUCTION

As recently as 1997, acute respiratory infections (ARI) were termed a “forgotten pandemic.”¹ These infections, which include pneumonia, bronchitis, bronchiolitis, otitis media, sinusitis, pharyngitis, laryngitis, measles, and pertussis, continue to cause 18% of all deaths worldwide among children younger than 5 years old and 8.2% of all disability and premature mortality.¹ Children living in developing countries have especially high morbidity and mortality rates. Despite the global prominence of acute lower respiratory infections, research investment in this area has lagged, with an inadequate proportion of all healthrelated research and development dollars being devoted to ARI. Recently, this situation has improved with the renewed concern about a global pandemic of influenza.¹

Although slow in coming, substantial gains in the battle against acute respiratory infections have been made by implementing a public health approach to prophylaxis and treatment. Major advances have occurred in understanding of the epidemiology and etiology of these infections. Data about risk factors continue to accumulate, with most information coming from the developed world, where funding, logistics, and infrastructure are available to support large, multidimensional epidemiologic studies related to ARI. Epidemiologic studies of the risk factors for lower respiratory tract infections have been pursued since the 1980s, with studies of nutrition, the effects of indoor and outdoor air pollution, and active and passive smoking being undertaken, and effective new vaccines being developed for the most important pathogens. In this chapter, the effects on public health, etiology, and risk factors of acute respiratory infections are reviewed, and the major methodological problems and requirements for future research are delineated.

IMPACT ON PUBLIC HEALTH

Developing Countries

Over the past 25 years, published data have indicated huge differences in ARI mortality rates between developing and developed countries. A World Health Organization (WHO) study in 1990 revealed that, in the world as a whole, the number of deaths attributable to ARI was 12 times greater in developing countries than in developed countries.² A recent WHO report estimated that in 2008 18% of all mortality in children aged younger than 5 years was attributable to acute respiratory infections (Table 19-1).³ The burden of mortality falls heaviest on developing countries. Those countries where infant mortality rates exceed 25 per 1000 suffer 98% of the world’s deaths from ARI in infants and 99% of those deaths in children aged 1 to 4 years.⁴ Furthermore, countries with infant mortality of approximately 100 per 1000 can be expected to contribute 58% of the ARI-related deaths in infants and 66% of those in children aged 1 to 4 years.⁴ As these estimates are based on national mortality reporting systems of varying quality, underestimation of the true mortality rate from ARI is possible.

Recent Estimates of Total Pneumonia Mortality and Morbidity

A systematic analysis of the global, regional, and national mortality in children younger than 5 years of age in 2008 was published recently.³ According to this analysis, an estimated 8.795 million deaths
occurred in children younger than 5 years in 2008. Of these deaths, 68% (5.970 million) were due to infectious diseases, with the largest proportion due to pneumonia (18%; 1.575 million), followed by diarrhea (15%; 1.335 million), and malaria (8%; 0.732 million). Neonatal deaths accounted for 41% of deaths and included preterm birth complications and deaths from a variety of other causes, including 386,000 deaths due to pneumonia in this age group.

Table 19-1 Causes of Mortality in Children Younger than 5 Years of Age in 2008

	Estimated Number (UR; millions)
Neonates aged 0-27 days
Preterm birth complications	1.033 (0.717-1.216)
Birth asphyxia	0.814 (0.563-0.997)
Sepsis	0.521 (0.356-0.735)
Other	0.409 (0.318-0.883)
Pneumonia^*	0.386 (0.264-0.545)
Congenital abnormalities^†	0.272 (0.205-0.384)
Diarrhoea^‡	0.079 (0.057-0.211)
Tetanus	0.059 (0.032-0.083)
Children aged 1-59 months
Diarrhoea^‡	1.257 (0.774-1.886)
Pneumonia^*	1.189 (0.789-1.415)
Other infections	0.753 (0.479-2.830)
Malaria	0.732 (0.601-0.851)
Other non-communicable diseases	0.228 (0.143-0.606)
Injury	0.279 (0.174-0.738)
AIDS^§	0.201 (0.186-0.215)
Pertussis^¶	0.195 (……)
Meningitis	0.164 (0.110-0.728)
Measles	0.118 (0.075-0.180)
Congenital abnormalities^†	0.104 (0.078-0.160)
Uncertainty range (UR) is defined as the 2.5-97.5 centile. ¨=data unavailable.
^*Estimated number of deaths in children younger than 5 years overall is 1.575 million (UR 1.046 million-1.874 million.) ^†Estimated number of deaths in children younger than 5 years overall is 0.376 million (UR 0.283 million-0.580 million). ^‡Estimated number of deaths in children younger than 5 years overall is 1.336 million (UR 0.822 million-2.004 million). ^§⁶Uncertainty range is based on UNAIDS’ estimated lower and upper bounds for deaths in children younger than 15 years. ^¶Crowcroft and colleagues’ sensitivity analysis presents extreme upper and lower values for various inputs.
Reprinted from The Lancet, 375, Black et al, Global, regional, and national causes of child mortality in 2008: a systematic analysis, 1969-87, Copyright 2010, with permission from Elsevier.

A previous systematic review of under-5 mortality in 2000 estimated the total global mortality in this age group to be 10.6 million, with 19% of these deaths due to pneumonia.¹ Estimates of the global burden of disease caused by infection with two major respiratory pathogens, Streptococcus pneumoniae and Haemophilus influenzae group B, have been published as well.⁵, ⁶ One study estimated that S. pneumoniae caused approximately 826,000 deaths (range: 582,000-926,000) in children aged 1-19 months in 2000.⁵ Of these deaths, 91,000 (range: 63,000-102,000) occurred in human immunodeficiency virus (HIV)-positive children and 735,000 (range: 519,000-926,000) in HIV-negative children; more than 61% of these deaths occurred in 10 African countries. S. pneumoniae infection was estimated to cause approximately 11% of all deaths in children 1-59 months of age. Acceleration of the use of protein-conjugated S. pneumoniae vaccines in these high-risk populations in Africa and Asia is a critical public health priority at present.

The global burden of disease of H. influenzae type b (Hib) in children 1-59 months of age was estimated to include approximately 8.13 million serious illnesses per year in 2000.⁶ Infections with this organism were estimated to have caused some 371,000 deaths (range: 247,000-527,000) in children 1-59 months of age in 2000. A highly effective and safe vaccine to prevent Hib infection is available, so these infections are almost entirely preventable. Fortunately, by 2006, 108 countries had implemented routine childhood vaccination with Hib vaccine.⁶

Global Alliance for Vaccines and Immunization

After the successful eradication of smallpox with a well-organized, coordinated public health program, which relied on an effective vaccine as a central component of the campaign, interest in using vaccines to control other infectious diseases increased. The first international vaccine effort after smallpox eradication in 1980 was the Childhood Vaccine
Initiative (CVI), which began in 1990 and functioned in collaboration with UNICEF and WHO’s Expanded Program on Immunization (EPI). The CVI was organized to help develop and provide preventive vaccines for developing country populations by providing a link between the pharmaceutical industry and public health authorities from developing country populations.⁷ Subsequently in 2000, the Global Alliance for Vaccines and Immunization (GAVI) was formed as a successor to CVI.

GAVI is a public-private partnership between the relevant experts and political decision makers in vaccines and immunization. It includes developing and donor countries, international development agencies and financial organizations (e.g., WHO, UNICEF and the World Bank), philanthropic organizations (i.e., the Bill and Melinda Gates Foundation), academia, the vaccine industry in both industrialized and developing countries, and representatives from civil society and the business community. GAVI was born and launched at the World Economic Forum in Davos in January 2000.

Developing countries can apply for funds from GAVI to support their immunization activities. Eligibility requirements include a gross domestic product per capita of less than $1500/year; a clear commitment to immunization, as shown by coverage of at least 50% children in each birth cohort with the six EPI vaccines; and a population of fewer than 150 million persons.⁸, ⁹ and ¹⁰

Over its first 10 years of operation, GAVI received and distributed $4.5 billion to process and buy vaccines and strengthen the health systems for 72 developing countries. A major success of this alliance has been a dramatic increase in the use of Hib vaccine. Use of this vaccine has resulted in an estimated 5.4 million future deaths prevented.⁹, ¹⁰ In addition, a vaccine to prevent meningococcal meningitis type A infections in the meningitis belt in Africa has been developed and distributed. Future challenges include the use of protein-conjugated polysaccharide vaccines for S. pneumoniae, typhoid fever and rotovirus vaccines, and other preventive vaccines.

Morbidity in Developing and Developed Countries

In contrast to the mortality data, ARI-related morbidity is similar for children in both developing and developed countries.⁴, ¹¹ In India, Ethiopia, and Costa Rica, the mean number of episodes of respiratory illness per year for children younger than 3 years of age was reported as 7.3, 7.9, and 4.9, respectively. In older children (3-5 years of age), the rates were also similar across India, Ethiopia, and Costa Rica—6.2, 6.6, and 5.7 episodes/year, respectively.¹¹, ¹², ¹³ and ¹⁴ Similarly, in the developed world, data from National Research Council studies of respiratory infection in young children in 12 countries suggest that the incidence of such disease in the first year of life ranges from 5 to 9 episodes.¹¹ Older data from the Seattle Virus Watch found that from 1965 to 1969, infants experienced a mean of 4.5 episodes per year.¹⁵ In Tecumseh, Michigan, infants experienced a mean of slightly more than 6 episodes of respiratory illness per year, and children aged 1 to 4 years experienced a mean of just more than 5 episodes per year.¹⁶ Caution is in order when considering these data, however: the studies from which these data were culled used widely differing methodologies and were reported several years ago. Thus prevalence rates might not be directly comparable or might have changed in recent years.

The incidence of chronic bronchitis (chronic obstructive pulmonary disease) is generally similar in developing and developed countries.¹⁷ For example, Lai et al. found 6.8% of an elderly group of Chinese living in Hong Kong to have chronic bronchitis,¹⁸ and pharmaceutical research indicated a 6.9% incidence in Taiwan.¹⁷ However, Pandey found a high incidence of chronic bronchitis (18%) in Nepal,¹⁹ with women and men equally affected by the disease. Indonesia also appears to have an especially high incidence of chronic bronchitis, although available data do not differentiate chronic bronchitis from acute bronchitis or account for multiple episodes during the observation period.¹⁷ An age profile in a sample of nearly 250,000 persons in Indonesia noted the incidence of chronic bronchitis was 15.7% in those persons aged 30 to 39 years, 19.3% in those aged 40 to 54 years, and approximately 6% in those aged 55 years or older.¹⁷

Developed Countries

In developed countries, ARI is the leading cause of morbidity, accounting for 20% of medical consultations, 30% of absences from work, and 75% of all antibiotic prescriptions.¹ Of all the ARIs, pneumonia has been the most thoroughly studied.

In 1900, pneumonia was the second leading cause of death in the United States, after tuberculosis.²⁰ Mortality rates per 100,000 population between 1910 and the present are shown in Figure 19-1.¹⁷ In 1918, pneumonia mortality skyrocketed due to complications from the great influenza pandemic, which that year killed more than 540,000 Americans. Since that time, the pneumonia-related mortality rate has fallen significantly because of better hygiene practices and the availability of effective treatment, including antipneumococcal serum, sulfa drugs, penicillin, and other antibiotics, as well as the use of influenza Hib and pneumococcal vaccines. The slight increase in
pneumonia-related deaths since 1990 primarily reflects the increase in proportion of Americans who are elderly and who have a high pneumonia risk.¹⁷ Mortality rates from influenza and pneumonia in the United States declined between 1999 and 2008, especially in persons older than age 75 and after 2003 (Table 19-2). This decline may be due in part to more widespread use of conjugated pneumocccal vaccines in children and more extensive use of influenza vaccines, in addition to the absence of a major influenza pandemic during these years. The sharp rise in influenza seen in the figure for 2009 reflects the severity of the 2009 influenza season. Currently, in the United States, mortality rates from community-acquired pneumonia range from 1% to 5% in outpatients and from 15% to 30% in inpatients, making it the eighth leading cause of death.²¹, ²², ²³ and ²⁴ Community-acquired pneumonia results in hospitalization in 20% of patients and 65 million days of restricted activity.²¹

Figure 19-1 Crude death rates for infectious diseases, United States 1990-1996. Reproduced from the Centers for Disease Control and Prevention (1999). MMWR, Vol. 48 No. 29 pp. 621-648.

Figure 19-2 pinpoints where in the United States deaths from pneumonia occurred between 1979 and 1992.²⁰ Although it was long theorized that a higher mortality rate from pneumonia occurred during winter and that cold temperatures promote pneumonia, these ideas are not consistent with the pattern of mortality.²⁰ Indeed, no definitive pattern with regard to climate is apparent in the United States from the 1979-1992 data: California had the highest pneumonia-related mortality rates; Georgia and Massachusetts had high rates; and North Dakota had the third lowest rate. The apparent contradiction raised by Florida, a state known as a retirement destination, being among those states with low rates of pneumonia is believed to be due to the “healthy retiree” effect. Healthy older people are able to retire to places such as Florida, whereas less healthy people in the older age group remain at home (Figure 19-2).²⁰

Hospital-acquired (nosocomial) pneumonia is the second most common nosocomial infection in the United States, but it is the type of nosocomial infection most frequently associated with a fatal outcome.²⁶ The annual incidence is 5 to 10 cases per 1000 hospital admissions, and up to 20 times this figure in patients on ventilators.²⁶ Mortality rates run as high as 33% to 50% in patients on ventilators.²⁷, ²⁸ The availability of penicillin, macrolides, and other antibiotics to treat pneumonia has greatly reduced the mortality and morbidity associated with this infection.

Age-specific incidence rates of minor episodes of respiratory illness (primarily upper respiratory tract infections) in the United States have varied little since 1933, as indicated by the results of studies that evaluated populations of varying compositions and used differing study methodologies and definitions of acute respiratory illness.²⁹ Over this period, however, pneumonia mortality rates have decreased in all age groups, except the elderly (Table 19-2). Indeed, since 1981, mortality rates have increased in pneumonia patients aged 55 years or older. It is unclear why this trend is occurring, although alterations in the types and pathogenicity of organisms causing pneumonia and changing host factors may be contributory factors.²⁹

Table 19-2 Influenza and Pneumonia Mortality Rates for the United States 1995-2009. A gradual decline in influenza and pneumonia related deaths is seen in the sparklines. The decline is accented by a sharp drop in pneumonia deaths following the introduction of the vaccine in 1999. There is also a recent increase in influenza in 2009 related to a more intense influenza season that year. Sparklines are simple graphical representations of data. Note they are not scaled relative to each other, but give a visual representation of the data in each age and disease category over the 15-year period.

Year	Age	All ages (age-adjusted)	18-24	25-44	45-54	55-64	65+ (age-adjusted)	65-74	75-84	85+
1995	Influenza and pneumonia	33.4	0.6	2.5	6.6	16.2	237.2	56.9	231.5	1020.9
1995	Influenza	0.2	*	*	*	0.1	1.6	0.4	1.2	8.2
1995	Pneumonia	33.2	0.6	2.5	6.5	16.1	235.5	56.5	230.3	1012.7
1996	Influenza and pneumonia	32.9	0.6	2.4	6.4	16.7	233.5	56.3	228.7	1002.1
1996	Influenza	0.3	*	0	0.1	0.2	2	0.4	1.6	9.6
1996	Pneumonia	32.6	0.6	2.4	6.3	16.6	231.5	55.8	227.1	992.5
1997	Influenza and pneumonia	33.3	0.7	2.3	6.5	17	236.3	56.4	231.6	1015.7
1997	Influenza	0.3	*	0	*	0.1	1.9	0.4	1.4	9.4
1997	Pneumonia	33	0.7	2.2	6.5	16.9	234.4	56	230.2	1006.3
1998	Influenza and pneumonia	34.6	0.7	2.3	6.2	16.8	247.4	59.3	240.1	1069.5
1998	Influenza	0.7	*	*	0.1	0.3	4.8	0.8	4.3	23
1998	Pneumonia	33.9	0.7	2.2	6.1	16.5	242.6	58.4	235.9	1046.5
1999	Influenza and pneumonia	23.5	0.5	1.6	4.6	11	167.4	37.2	157	751.8
1999	Influenza	0.6	*	0	0.1	0.3	4.4	0.9	3.7	21.2
1999	Pneumonia	22.9	0.5	1.6	4.6	10.7	163	36.4	153.3	730.6
2000	Influenza and pneumonia	23.7	0.6	1.7	4.7	11.9	168.6	39.1	160.3	744.1
2000	Influenza	0.6	*	0	0.1	0.4	4.5	1.1	4	20.2
2000	Pneumonia	23.1	0.6	1.6	4.6	11.5	164.2	38	156.4	723.9
2001	Influenza and pneumonia	22	0.5	1.6	4.6	10.7	155.8	36.3	148.5	685.6
2001	Influenza	0.1	*	*	*	0.1	0.5	0.1	0.4	2.2
2001	Pneumonia	21.9	0.5	1.5	4.6	10.6	155.3	36.2	148.1	683.4
2002	Influenza and pneumonia	22.6	0.5	1.6	4.8	11.2	160.7	37.5	156.9	696.6
2002	Influenza	0.2	*	*	*	0.1	1.7	0.3	1.4	8.8
2002	Pneumonia	22.4	0.4	1.5	4.7	11.1	159	37.2	155.5	687.8
2003	Influenza and pneumonia	22	0.6	1.6	5.2	11.2	154.8	37.3	151.1	666.1
2003	Influenza	0.6	0.1	0	0.1	0.3	3.9	1	3.6	17
2003	Pneumonia	21.4	0.6	1.6	5.1	10.9	150.9	36.3	147.5	649.1
2004	Influenza and pneumonia	19.8	0.5	1.4	4.6	10.8	139	34.6	139.3	582.6
2004	Influenza	0.4	*	0	0.1	0.2	2.5	0.5	2.2	11.4
2004	Pneumonia	19.4	0.5	1.4	4.5	10.7	136.5	34	137	571.2
2005	Influenza and pneumonia	20.3	0.5	1.5	5.1	11.3	141.9	35.5	142.2	593.9
2005	Influenza	0.6	*	0	0.1	0.2	4.1	0.7	3.7	19.3
2005	Pneumonia	19.7	0.4	1.5	5	11.1	137.8	34.8	138.5	574.7
2006	Influenza and pneumonia	17.8	0.5	1.4	4.6	10	123.7	32	127.8	502.5
2006	Influenza	0.3	*	*	0.1	0.1	1.7	0.4	1.8	7
2006	Pneumonia	17.5	0.5	1.4	4.6	9.8	122	31.6	125.9	495.5
2007	Influenza and pneumonia	16.2	0.5	1.3	4.4	9.6	112.3	28.7	114.1	463.2
2007	Influenza	0.1	*	0	*	0.1	0.6	0.2	0.6	2.5
2007	Pneumonia	16.1	0.4	1.3	4.3	9.5	111.7	28.5	113.5	460.7
2008	Influenza and pneumonia	16.9	0.6	1.5	5.1	11.1	115.6	31.1	119.1	465.2
2008	Influenza	0.5	*	0.1	0.2	0.4	3.2	0.6	3.2	14.3
2008	Pneumonia	16.4	0.5	1.4	4.9	10.7	112.3	30.4	115.9	450.9
2009	Influenza and pneumonia	16.2	1.1	2.6	6.5	11.9	104	30.1	105.9	413.5
2009	Influenza	0.9	0.5	0.8	1.5	1.4	1.6	1.1	1.6	3.6
2009	Pneumonia	15.3	0.6	1.7	5	10.5	102.5	29	104.4	409.8

Data from the Centers for Disease Control and Prevention (2012). National Center for Health Statistics. http://www.cdc.gov/nchs/. Last updated November 8, 2012. Accessed November 13, 2012.

Figure 19-2 Influenza and pneumonia death rate per 100,000. “Number of Deaths per 100,000 Population Caused by Influenza and Pneumonia, 2009”, statehealthfacts.org, The Henry J. Kaiser Family Foundation, May 2009.

Despite decreasing mortality rates for pneumonia and influenza in young children and infants, lower respiratory tract illness (croup, bronchitis, bronchiolitis, and pneumonia) remains an important cause of morbidity in this age cohort, annually affecting approximately 25% of children aged less than 1 year and 18% of children aged 1 to 4 years.³⁰, ³¹, ³², ³³, ³⁴ and ³⁵ In 1989, Wright et al. reported the cumulative incidence of first episodes of lower respiratory tract infection in infants to be 32.9% in children of families participating in a prepaid health plan.³⁶ This higher rate might have been seen because of differing diagnostic criteria compared with previous studies,³², ³³ and ³⁴ an active rather than a passive followup regimen, and increased physician attendance because of lack of a financial disincentive in a prepaid health plan. In older children and adults in the United Kingdom, ARI accounted for almost one fourth of all
primary care contacts, and one third of days taken off work.³⁷ In a survey in Australia, 17% of patients aged 15 years or older and 43% of children aged less than 15 years consulted a doctor for respiratory symptoms during the 2 weeks preceding the survey.³⁸ A study in Adelaide, South Australia, showed that children younger than 5 years experienced a mean of seven episodes of respiratory illness per year, which prompted three doctor visits, 15 days of medication, and 52 days of respiratory symptoms annually.³⁸, ³⁹

The incidence of chronic bronchitis is similar in the United States and Europe.²⁰ In 1994, Enright et al. reported that 5.1% to 5.4% of the middle-aged to elderly population in the United States have chronic bronchitis, with a lower prevalence in nonsmokers.⁴⁰ In Europe, chronic bronchitis has been reported to affect 3.7% of people in Denmark,⁴¹ 4.5% in Norway,⁴² 6% and 6.4% in Barcelona and Valencia, Spain, respectively,⁴⁴ and 6.7% in Sweden.⁴⁵

Table 19-3 Classification of Acute Respiratory Infection Clinical Syndromes

	Case-Management Classification (children aged 2 months to 4 years)
Stridor	Wheezing	No wheezing	“Traditional” Classification
Mild	Mild	Mild	Upper respiratory tract syndromes
Hoarseness plus “barking” cough; no stridor when calm = mild croup	Improves with bronchodilator; respiratory rate <50/min = mild bronchiolitis or asthma	Cough; nasal obstruction; respiratory rate <50/min = URI,^* cold	Common cold; URI^*
			Acute otitis media
			Pharyngitis/tonsillitis
			Acute sinusitis
Home care; no antibiotic	Oral salbutamol	No antibiotic; home care
	Moderate	Moderate	Middle respiratory tract syndromes
	Improves with bronchodilator; respiratory rate 50-70/min = mild bronchiolitis or asthma	Respiratory rate >50/min; no chest indrawing = pneumonia	Croup Laryngotracheobronchitis
				Epiglottitis
	Consider antibiotic; oral salbutamol; home care	Antibiotic; home care		Laryngitis
				Tracheitis
Severe	Severe	Severe	Lower respiratory tract syndromes
Stridor when calm; chest indrawing = severe croup or epiglottitis	No improvement with bronchodilator; respiratory rate >70/min = severe bronchiolitis or asthma	Respiratory rate >50/min; chest indrawing = severe pneumonia		Bronchiolitis
				Bronchitis
				Pneumonia
Admit; antibiotic; manage airway	Admit; bronchodilators; consider oxygen and antibiotics	Admit; antibiotic
	Very severe	Very severe
	Cyanosis or inability to drink = very severe bronchiolitis or asthma	Cyanosis or inability to drink = very severe pneumonia
	Admit; bronchodilators; oxygen and consider antibiotic	Antibiotic; admit; oxygen
^*URI, upper respiratory infection.
Reproduced from NMH Graham, The Epidemiology of Acute Respiratory Infection in Children and Adults: A Global Perspective, Epidemiologic Reviews, Vol. 12, p. 152 , Table 2. © 1990. By permission of Oxford University Press.

In the United States in 1998, community-acquired pneumonia accounted for an estimated $3.6 billion in expenditures for treating patients younger than age 65 years and $4.8 billion for treating patients aged 65 years or older.⁴⁶ In 1999, costs associated with acute exacerbations of chronic bronchitis were $419 million in patients younger than 65 years and three times that much ($1.2 billion) for patients aged 65 years and older.⁴⁷

CLASSIFICATION OF ACUTE RESPIRATORY INFECTIONS

Two basic systems are most commonly used to classify acute respiratory infections: the case-management classification system and the “traditional” clinical classification system (Table 19-3).

Case-Management Classification

In an effort to reduce pneumonia-related mortality, WHO has developed a simple case-management approach to be followed by village healthcare workers. Before 1988, simplified case classifications, as used in India,⁴⁸ Papua New Guinea,⁴⁹ and other developing countries,⁵⁰ were applied successfully by health workers to determine when children should be given antibiotics or referred to secondary or tertiary level care. Adherence with these classifications resulted in reductions in both pneumonia-related mortality and in overall mortality. The simplified case definitions were based on respiratory rates recorded among children with symptoms of respiratory infections. The sensitivity and specificity of increased respiratory rate were determined using clinically confirmed or chest radiograph-confirmed pneumonia (or “lower respiratory tract infection”) as the putative gold standard. For example, using receiver operating characteristic (ROC) curves, Cherian et al. showed that higher respiratory rates were more sensitive and specific for infants than for older children aged less than 5 years (Figure 19-3).⁵¹

In 1988, changes were made to improve the specificity of case-management guidelines.⁵⁰, ⁵², ⁵³, ⁵⁴ and ⁵⁵ Although the primary focus remained on pneumonia, classification and management of syndromes causing stridor and wheezing were directly addressed, and otitis media was classified separately (Table 19-3). Other major differences from the earlier classification were in the younger-than-2-month-old age group, for whom the newer classification recognized that a respiratory rate greater than 50 breaths/minute and some chest in-drawing could be considered normal⁵²; with regard to cough, which is often absent in neonates with pneumonia, the newer classification specified that this sign is not sufficiently sensitive to be used as an indicator of severe disease. The newer case-management classification recommended that signs of general sepsis should be sought (i.e., fever, feeding problems, drowsiness, convulsions, abdominal distention, hypothermia) as well as more specific signs of pneumonia (respiratory rate greater than 60 breaths/minute, severe chest in-drawing, respiratory grunt) when determining whether to prescribe antibiotics and to hospitalize patients.⁵²

Figure 19-3 Receiver operating characterisitcs (ROC) curves for respiratory rates as indicators of lower respiratory tract infections in infants and children. Reprinted from The Lancet, 322, Cherian et al. Evaluation of simple clinical signs for the diagnosis of acute lower respiratory tract infection, 126, Copyright 1988, with permission from Elsevier.

Clinical Classification

Acute respiratory infections can be classified by the site of primary pathology (Table 19-3). This system is preferred by most physicians, and it is compatible with the International Classification of Diseases
(ICD) system. However, such classification can lead to some confusion because infections are not always limited to one part of the respiratory tract. Moreover, clinicians may disagree on whether an acute respiratory infection can be termed “upper,” “middle,” or “lower.” Stridor-causing conditions often have been classified as upper respiratory tract infections, thus suggesting mild disease, which can be a misleading assumption. Because stridor can cause severe, and possibly fatal, respiratory distress, consideration has been given to classifying stridor-causing conditions as acute lower respiratory infections.⁵⁵

PATHOGENS RESPONSIBLE FOR ACUTE RESPIRATORY INFECTIONS

In a large percentage of patients with acute respiratory infections, the pathogens responsible for infection are not known. This problem is especially notable with regard to community-acquired pneumonia, for which the causative organism is unknown in approximately 98% of those individuals treated as outpatients and 50% to 60% of those persons treated as hospital inpatients.⁵⁶, ⁵⁷

Viruses

Many upper respiratory viral infections in children and adults are caused by rhinoviruses (30% to 50%) or coronaviruses (5% to 20%), with the remainder (30% to 65%) due to influenza virus, parainfluenza virus, respiratory syncytial virus, adenoviruses, and certain enteroviruses.⁵⁸ These infections are generally mild, are self-limiting, and do not involve respiratory distress.

In children, viral causes of the acute lower respiratory tract infections, pneumonia, bronchiolitis, and croup generally appear to be similar in both developed and developing countries.⁵⁹, ⁶⁰, ⁶¹, ⁶², ⁶³, ⁶⁴, ⁶⁵, ⁶⁶, ⁶⁷, ⁶⁸ and ⁶⁹ The primary pathogens are respiratory syncytial virus; parainfluenza virus types 1, 2, and 3; influenza virus types A and B; adenoviruses; metapneumovirus; and enteroviruses. Respiratory syncytial virus is most commonly associated with bronchiolitis, and parainfluenza virus (especially type 1) is more often associated with croup.⁷⁰, ⁷¹, ⁷² and ⁷³ In developing countries, measles contributes to croup and serious morbidity in other lower respiratory infections more than it does in the developed world.⁶⁹, ⁷⁴, ⁷⁵ Approximately 90% of cases of acute bronchitis are caused by viruses, including influenza virus, parainfluenza virus, and rhinovirus.⁶⁸, ⁷⁶

In adults, viral causes of pneumonia are generally less important than nonviral causes. Nevertheless, influenza has been associated with a significant proportion of cases in adults, perhaps causing as many as half of all virus-associated cases and approximately 8% to 10% of all pneumonias in general.³², ⁷⁹, ⁸⁰ Respiratory syncytial virus and parainfluenza also have been identified in some adult cases of pneumonia, but these viruses are less common than influenza.⁸⁰ In developed countries, influenza epidemics in elderly patients (older than 65 years of age) are associated with especially high mortality and hospitalization rates from acute respiratory infections; thus influenza poses a major health risk to this age group.⁷⁹ More than 50% of excess hospitalizations and more than 80% of influenza-related deaths occur in persons aged 65 years and older, in whom mortality is 30 to 50 times greater than in younger adults and adolescents.⁷⁹, ⁸⁰ In developing countries, viral causes of adult lower respiratory tract infections have not been widely studied. No indication is seen that viral pathogens are implicated to any greater degree in pneumonia cases in developing countries than they are in developed countries, although the mode of transmission in the poorer nations (unwashed hands, contaminated water) can differ from those in the developed world (daycare facilities, contaminated aerosols).⁷⁶, ⁸⁰

Several viruses have been identified recently to be important causes of acute respiratory infection. The severe acute respiratory syndrome (SARS) coronavirus caused a major pandemic when it emerged in southern China in 2002. This virus is a new human pathogen that crossed species to infect humans—that is, a zoonotic infection. While no new cases of SARS have been seen since this epidemic, its emergence reminds us of the potential for zoonosis and the risk of previously unknown viruses infecting humans. The epidemiology of the SARS pandemic is covered in detail in the chapter on emerging infections.

In 2001, investigators in the Netherlands isolated a new virus from children and adults with acute respiratory tract infection.⁸¹ This RNA virus is closely related to avian pneumovirus. In the last few years, metapneumovirus has been isolated from patients with acute respiratory infection in the United States, Australia, Canada, and the United Kingdom.⁸², ⁸³, ⁸⁴, ⁸⁵ and ⁸⁶ A study was published recently from Vanderbilt University in Nashville, Tennessee, in which nasal washes were collected prospectively from 463 infants and children who were seen for acute lower respiratory infection between 1976 and 2001.⁸⁷ A viral cause other than metapneumovirus was detected in 41% of these children. Of 248 specimens available for which no other pathogen was detected, 49 (20%) contained metapneumovirus⁸⁶; 28% of illnesses occurred
between December and April; and 2% of infected children were hospitalized. The seasonal distribution of metapneumoviruses and other viral pathogens in this study is shown in Figure 19-4.

Figure 19-4 Epidemiologic pattern of lower respiratory tract infections with human metapneumovirus and other viruses. From Williams et al. Human metapneumovirus and lower respiratory tract disease in otherwise healthy infants and children. N Engl J Med. Vol. 350(5):446, Figure 1. © 2004 Massachusetts Medical Society. Reprinted with permission from Massachusetts Medical Society.

In 2005, a human parvovirus was identified in pooled respiratory secretions using host DNA depletion, random polymerase chain reaction (PCR) amplification, sequencing, and bioinformatics.⁸⁷ This newly identified virus was named human bocavirus (HBoV). It was subsequently identified in respiratory secretions of 17 additional children with clinical lower respiratory tract infection. Following this initial report, investigators from countries around the world have identified HBoV in respiratory secretion of 5% to 10% of children with pneumonia.⁸⁸, ⁸⁹, ⁹⁰, ⁹¹ and ⁹² In addition, HBoV has been identified in fecal samples of children with gastroenteritis.⁹³, ⁹⁴

Sinnombre virus is another recently identified virus that is responsible for lower respiratory infections during epidemic or endemic transmission from its reservoir in rodents. The epidemiology of the hantavirus pulmonary syndrome associated with sinnombre virus infections is discussed in the chapter on emerging infections.

Prior to the use of molecular diagnostic methods to identify the agents causing acute respiratory tract infection, most epidemiologic studies had been hampered by the limited laboratory techniques available
to isolate viral pathogens. Isolation of virus from 20% to 25% of specimens has been the maximal rate in several well-conducted studies.³¹, ⁷⁶, ⁹⁵, ⁹⁶, ⁹⁷, ⁹⁸ and ⁹⁹ Recent improvements in these techniques might be expected to result in higher isolation rates becoming more common. When combinations of viral culture and immunofluorescence techniques were used in one study of lower respiratory infections in infants to identify viral pathogens from throat and nasopharyngeal swabs, a 66% isolation rate was reported.³²

The use of nucleic acid amplification methods to identify respiratory viruses has increased the proportion of illnesses in which a viral pathogen has been identified to substantially more than 50%.⁸⁰, ⁸¹

Bacteria

Microbiologic isolation of the specific pathogen that is responsible for community-acquired pneumonia is difficult and is rarely done in current medical practice. Expectorated sputum is commonly contaminated with bacterial flora from the upper airway.⁹⁸ Examination of a Gram-stained sputum is very rarely done at present because of the belief that it is an insensitive test.⁹⁹ In addition, the Clinical Laboratory Integrity Act (CLIA) regulations have discouraged the routine microscopic examination of a stained sputum because they require such examinations be performed only by a licensed medical technician.⁹⁹ Sputum obtained by transtracheal aspiration or bronchoscopy is more likely to identify the infecting pathogen, but these procedures carry some risks for the patients, are more difficult to implement, and may be susceptible to some contamination. Percutaneous needle aspirates of the lung have been reported to be a sensitive method to identify the pathogen in patients with lobar pneumonia.¹⁰⁰, ¹⁰¹

Whereas the specific pathogen responsible for an adult patient with community-acquired pneumonia (CAP) was identified in 50% to 80% of cases prior to the mid-1980s, it is identified in less than 10% of cases in current medical practice.⁹⁹ This trend has implications for treatment. A prospective study was done comparing the outcome (i.e., mortality, length of hospital stay, resolution of fever, and clinical treatment failure) in 262 patients who were randomized to receive either empirical broad-spectrum antibiotic treatment to treat the major pathogens in pneumonia cases or initial identification of the infecting organism followed by pathogen-directed treatment¹⁰²; this study found no difference in these outcomes between the two groups of patients. Another study compared the outcome of 18,209 Medicare patients who were hospitalized for CAP according to when antibiotic treatment was started after hospital admission.¹⁰³ This study found a reduction in in-hospital mortality and length of stay among patients in whom antibiotics were begun within 4 hours of admission.¹⁰³ Initiation of antibiotic therapy for CAP has been utilized subsequently as one of the measures of quality of care of hospitalized patients with pneumonia.¹⁰⁴ As rapid antibiotic use is now considered a measure of quality care for this indication, health providers have been pressured to start therapy as quickly as possible after pneumonia diagnoses, promoting the use of broad-spectrum empirical therapy over the strategy of first identifying the infecting organism and then tailoring the therapy for the causative pathogen.¹⁰⁴ On the one hand, this practice has resulted in some unnecessary use of some antibiotics in patients with CAP, thereby increasing the risk of the development of antibiotic resistance strains of pathogens. On the other hand, earlier therapy may improve the outcome for some patients.

One consideration in selecting the therapy for CAP patients is whether they might be infected with an “atypical agent,” such as Mycoplasma pneumoniae hominis, Legionella pneumophila, Chlamydia pneumoniae, or a viral pathogen. Of these difficult-to-diagnose microorganisms by culture or a stained sputum exam, Legionella may be the most significant because it requires specific therapy to avoid significant morbidity or mortality. Legionella should be considered if the patient acquired pneumonia on a cruise ship or in a hotel.¹⁰⁵ A recent outbreak of Legionella infection was reported among persons exposed to the pathogen from a decorative “water wall.”¹⁰⁶ These waterfalls had been installed in hospital waiting areas as attractive and relaxing features; however, it is likely most will now be removed to prevent Legionnaires’ disease. Rapid antigen detection tests have been approved by the Food and Drug Administration (FDA) to detect S. pneumoniae or Legionella pneumonphila type 1 antigens in the urine.¹⁰⁶, ¹⁰⁸ In addition, a multiplex reverse transcription-polymerase chain reaction (RT-PCR) assay to detect 12 viral pathogens simultaneously in patients with CAP has been licensed.¹⁰⁹ The more frequent use of these assays in the future should facilitate more rapid initiation of effective therapy in patients with acute respiratory infection and improve outcomes.

Active Bacterial Core Surveillance of the Emerging Infections Program Network

To better understand the epidemiology of several important bacterial pathogens that cause acute community-acquired lower respiratory infections, sepsis, and meningitis, the Centers for Disease Control and Prevention (CDC) established the Active Bacterial Core (ABC) Surveillance Program in 1995.¹¹⁰ This program collects and analyzes microbiologic and
molecular data on five organisms that are isolated in cultures from usually sterile sites (i.e., blood, cerebrospinal fluid, joint fluid) by 600 laboratories in all or part of seven states that are participating in the CDC-funded Emerging Infections Program (EIP). The pathogens that are evaluated include S. pneumoniae, H. influenzae, N. meningitides, group A Streptococcus (S. pyogenes) and group B Streptococcus.

Data collected in the ABC study have allowed evaluation of the effectiveness of three vaccines: pneumococcal, H. influenzae type b, and meningococcal. The emergence of serotypes not contained in the S. pneumoniae and N. meningitis vaccines and the antibiotic sensitivity patterns of these important pathogens also have been evaluated in the ABC program. An estimated 6% to 10% of untreated pulmonary infections involving S. pneumoniae or H. influenzae result in sepsis or meningitis, so the ABC system will identify only a minority of pneumonia caused by these pathogens. Nevertheless, many of the more serious infections will be detected.

Developing Countries

Bacteria-caused lower respiratory infections are more common in children in developing countries than in the developed world, where viral infections are more often encountered.³³ Reliable data regarding the pathogens responsible for pneumonia in adults in developing countries have come chiefly from hospital studies. Streptococcus pneumoniae is by far the most important cause of pneumonia in adults: it is associated with as many as 70% of cases in which a pathogen is isolated,¹¹¹ with Haemophilus influenzae and Staphylococcus aureus being relatively less important in adults than in children (together accounting for approximately 10% of cases). Streptococcus pyogenes and Corynebacterium diphtheriae most commonly cause pharyngitis and tonsillitis. Acute epiglottitis is caused chiefly by H. influenzae type b, and whooping cough by Bordetella pertussis.¹¹², ¹¹³ In otitis media, S. pneumoniae and H. influenzae are the most commonly isolated bacteria,¹¹³, ¹¹⁴ but Moraxella catarrhalis has been isolated in 27% of patients in some series.¹¹⁵ S. pneumoniae and H. influenzae are also important in acute sinusitis in children¹¹⁶, ¹¹⁷ and in adults.¹¹⁸

Table 19-4 Microbiological Characteristics of Community-Acquired Pneumonia (CAP)

	Prevalence, %
Origin	North America^a	British Thoracic Society^b
Streptococcus pneumoniae	20-60	60-75
Haemophilus influenzae	3-10	4-5
Staphylococcus aureus	3-5	1-5
Enterobacteriaceae	3-10	Rare
Legionella	2-8	2-5
Mycoplasma pneumoniae	1-6	5-18
Chlamydophila pneumoniae	4-6	–
Aspiration	6-10	–
Viruses	2-15	8-16
Note: Data from Mundy et al. [8].
^aBased on 15 reports from North America. ^bBased on an analysis of 453 adults in a prospective study of CAP in 25 British hospitals. Ellipses indicate that no studies were performed to detect the designated agent.
Reproduced from Bartlett (2011). Diagnostic tests for agents of Community Acquired Pneumonia, Clinical Infectious Diseases 2011:52(S4);S296-S304.

Berman and McIntosh¹¹⁹ have reviewed studies of bacteria isolated from children in developing countries using lung aspiration techniques. The most frequently isolated pathogens were H. influenzae, S. pneumoniae (together accounting for 54% of isolates), and S. aureus (accounting for 17% of isolates). A subsequent study in Zimbabwe found a similar pattern of pathogens responsible for lower respiratory infections.¹²⁰

Developed Countries

The key bacterial pathogens responsible for CAP in the developed world appear to be the same as those in developing countries. A literature review of 15 published reports from North America showed the most common bacterial pathogens to be S. pneumoniae (20% to 60% of isolates) and H. influenzae (3% to 10%) (Table 19-4). A meta-analysis of 122 published studies between 1966 and 1995 (N = 7057) indicated that S. pneumoniae was responsible for 66% of
deaths in CAP cases.¹²¹ A review of health care associated pneumonia indicated that gram-negative bacteria account for 50% to 70% of cases (Table 19-5).¹²¹ The most frequently isolated pathogen is Pseudomonas aeruginosa, followed by a diverse array of Enterobacteriaceae. These bacteria reach the lower airways by aspiration of gastric contents. Patients being treated with ventilators are at the highest risk of health care associated pneumonia.

Table 19-5 Pathogens Most Commonly Involved with Health Care Associated Pneumonia

	Literature	The British Thoracic	Meta-analysis^‡
Microbial Agents	*Review^ (%)**	Society^† (%)	Cases (%)	Deaths (%)
Bacteria
-Streptococcus pneumoniae	20-60	60-75	65	66
-Haemophilus influenzae	3-10	4-5	12	7
-Staphylococcus aureus	3-5	1-5	2	6
-Gram-negative bacilli	3-10	Rare	1	3
-Miscellaneous agents^§	3-5	(Not included)	4	9
Atypical pathogens	10-20	—	12	6
-Legionella sp.	2-8	2-5	4	5
-Mycoplasma pneumoniae	1-6	5-18	7	1
-Chlamydia pneumoniae	4-6	(Not included)	1	<1
Viral	2-15	8-16	3	≤1
Aspiration pneumonia	6-10	(Not included)	—	—
No diagnosis	30-60	—	—	—
^Based on analysis of 15 published reports from North America.⁹⁸ Low and high values are deleted. ^†Estimates are based on analysis of 453 adults in prospective study of community-acquired pneumonia in 25 British hospitals.⁹⁸ ^‡Meta-analysis of 122 published studies of community-acquired pneumonia in the English language literature 1966 to 1995; data are limited to 7057 patients who had an etiologic diagnosis.⁹⁸ Percentage of death column refers to percentage of all deaths attributed to the designated pathogen. ^§Includes Moraxella catarrhalis, group A streptococcus, and Neisseria meningitidis* (each 1% to 2%).
J.G. Bartlett (1998). Approach to the Patient with Pneumonia. Ed. Gorbach, Infectious Disease, 2nd edition. pp. 553-564 . By permission of Oxford University Press.

Approximately 50% to 75% of infective exacerbations of chronic bronchitis are bacterial in origin.¹⁷ Studies conducted in the Northern Hemisphere have consistently shown H. influenzae to be the major pathogen and M. catarrhalis to be the second most common pathogen.⁷⁶

Other Pathogens

Mycoplasma pneumoniae, Chlamydia species, Legionella species, and Pneumocystis carinii are the nonviral respiratory pathogens most frequently responsible for pneumonia and acute bronchitis in both children and adults.¹²², ¹²³ and ¹²⁴

M. pneumoniae can also cause upper respiratory infections.¹²², ¹²³, ¹²⁵ Of the Chlamydia species, C. trachomatis is implicated more in cases of pneumonia in young infants,¹²⁶ and C. pneumoniae (also called TWAR, based on the names of the first two isolates, TW-183 and AR-39) is responsible primarily for pneumonia cases among older children and adults.¹²⁷, ¹²⁸ Most of the cases of P. carinii pneumonia that have been reported since the early 1980s have occurred in patients with acquired immunodeficiency syndrome (AIDS).¹²⁴

P. carinii pneumonia occurs as an opportunistic infection in patients with AIDS most frequently when their CD4⁺ lymphocyte counts fall below 200 cells/mm³. Approximately 60% to 80% of AIDS patients will develop P. carinii pneumonia at some stage during the course of their illness.¹²⁴ In developed countries, both children and adults with AIDS and other conditions associated with immunosuppression are at significant risk of P. carinii pneumonia. In countries in sub-Saharan Africa, AIDS-associated P. carinii pneumonia is less common, but the reasons for this lower incidence are unclear.¹²⁴ HIV-infected patients also have a greater incidence of CAP caused by the intracellular bacterial pathogens Salmonella and Legionella than patients not infected with HIV.¹²⁴

RISK FACTORS

The risk factors for community-acquired pneumonia and hospital acquired pneumonia are summarized in Table 19-6 and Table 19-7, respectively.¹¹⁰ Specifics pertaining to key risk factors are provided in this section.

Table 19-6 Acute Repiratory Infection Morbidity Rates in Four U.S. Cohort Studies

			Mean Incidence per Year by Age Group (years)
Investigators	Years	Population	<1	1-2	3-4	5-9	10-14	15-19	20-24	25-29	30-39	40-49	50-59	≥60
van Volkenburgh and Frost²⁵	1924	Public health service families		3.0		2.7	1.9	1.4		2.0^*	1.8^†	1.6^‡	1.4^§
	1928-1929	Baltimore, MD, families		4.5		3.5	3.5	2.4	2.8		2.7	2.4	1.7
	1929-1930	Baltimore, MD, families		4.5		3.8	2.8	2.3	2.7		2.6	2.3	2.1
Gwaltney et al.²⁶	1963-1966	Insurance company employees						2.5		2.1	2.2	1.7
Fox et al.²⁷	1965-1969	Seattle, WA, families	5.1	5.8	5.8	3.8	2.3
Monto and Ullman¹³	1969-1971	Tecumseh, MI, families	6.1	5.7	4.7	3.5	2.7	2.4	2.8	2.7	2.3	1.7	1.6	1.3
^*25-34 years of age. ^†35-44 years of age. ^‡45-54 years of age. ^§≥55 years of age.
^¶≥45 years of age.
Reproduced from NMH Graham, The Epidemiology of Acute Respiratory Infection in Children and Adults: A Global Perspective, Epidemiologic Reviews, Vol. 12, pp. 149-178 , © 1990, by permission of Oxford University Press.