In the late nineteenth century, pairs of micrococci were first recognized in the blood of rabbits injected with human saliva by both Louis Pasteur working in France and George Sternberg, an American army physician. The important role of these micrococci in human disease was not appreciated at that time. By 1886, when the organism was designated “pneumokokkus” and Diplococcus pneumoniae, the pneumococcus had been isolated by many independent investigators, and its role in the etiology of pneumonia was well known. In the 1930s, pneumonia was the third leading cause of death in the United States (after heart disease and cancer) and was responsible for ∼7% of all deaths both in the United States and in Europe. While pneumonia was caused by a host of pathogens, lobar pneumonia—a pattern more likely to be caused by the pneumococcus—accounted for approximately one-half of all pneumonia deaths in the United States in 1929. In 1974, the organism was reclassified as Streptococcus pneumoniae.
Pneumococci are spherical gram-positive bacteria of the genus Streptococcus. Within this genus, cell division occurs along a single axis, and bacteria grow in chains or pairs—hence the name Streptococcus, from the Greek streptos, meaning “twisted,” and kokkos, meaning “berry.” At least 22 streptococcal species are recognized and are divided further into groups based on their hemolytic properties. S. pneumoniae belongs to the α-hemolytic group that characteristically produces a greenish color on blood agar because of the reduction of iron in hemoglobin (Fig. 134-1). The bacteria are fastidious and grow best in 5% CO2 but require a source of catalase (e.g., blood) for growth on agar plates, where they develop mucoid (smooth/shiny) colonies. Pneumococci without a capsule produce colonies with a rough surface. Unlike that of other α-hemolytic streptococci, their growth is inhibited in the presence of optochin (ethyl hydrocuprein hydrochloride), and they are bile soluble.
Pneumococci growing on blood agar, illustrating α hemolysis and optochin sensitivity (zone around optochin disk). Inset: Gram's stain, illustrating gram-positive diplococci. (Photographs courtesy of Paul Turner, Shoklo Malaria Research Unit, Thailand.)
In common with other gram-positive bacteria, pneumococci have a cell membrane beneath a cell wall, which in turn is covered by a polysaccharide capsule. Pneumococci are divided into serogroups or serotypes based on capsular polysaccharide structure, as distinguished with rabbit polyclonal antisera; capsules swell in the presence of specific antiserum (the Quellung reaction). The 91st and 92nd serotypes, 6C and 6D, have most recently been identified with monoclonal antibodies and by serologic, genetic, and biochemical means, respectively. Within the 92 serotypes there are 21 serogroups, each containing two to five serotypes with closely related capsules. The capsule protects the bacteria from phagocytosis by host cells in the absence of type-specific antibody and is arguably the most important determinant of pneumococcal virulence. Unencapsulated variants tend not to cause invasive disease.
Within the cytoplasm, cell membrane, and cell wall, many molecules that may play a role in pneumococcal pathogenesis and virulence have been identified (Fig. 134-2). These proteins are often involved in direct interactions with host tissues or in concealment of the bacterial surface from host defense mechanisms. Pneumolysin is a secreted cytotoxin thought to result in cytolysis of cells and tissues, and LytA enhances pathogenesis. A number of cell wall proteins interfere with the complement pathway, thus inhibiting complement deposition and preventing lysis and/or opsonophagocytosis. The pneumococcal H inhibitor (Hic) impedes the formation of C3 convertase, while pneumococcal surface protein C (PspC), also known as choline-binding protein A (CbpA), binds factor H and is thought to accelerate the breakdown of C3. PspA and CbpA inhibit the deposition of or degrade C3b. The numerous pneumococcal proteins thought to be involved in adhesion include the ubiquitous surface-anchored sialidase (neuraminidase) NanA, which cleaves sialic acid on host cells and proteins, and pneumococcal surface adhesin A (PsaA). Pili recently recognized by electron microscopy may also play an important role in binding to cells. Some of the antigens mentioned above are potential vaccine candidates (see “Prevention,” below).
Schematic diagram of the pneumococcal cell surface, with key antigens and their roles highlighted.
Although the capsule surrounding the cell wall of S. pneumoniae is the basis for categorization by serotype, the behavior and pathogenic potential of a serotype may also be related to the genetic origin of the strain. Molecular typing is therefore of considerable interest. Initially, techniques such as pulsed-field gel electrophoresis were used to determine genetic relatedness; such techniques have been superseded by sequencing of housekeeping genes to define a clone (multilocus sequence typing, MLST). For S. pneumoniae, alleles at each of the loci aroE, gdh, gki, recP, spi, xpt, and ddl are sequenced and compared with all of the known alleles at that locus. Sequences identical to a known allele are assigned the same allele number, whereas those differing from any known allele—even at a single nucleotide site—are assigned new numbers. Software for assignment of alleles at each locus is available on the pneumococcal MLST web site (http://spneumoniae.mlst.net/), and the allelic profile of each isolate and its consequent sequence type are generated. With the advent of high-throughput and relatively inexpensive sequencing techniques, whole-genome sequencing will soon supersede MLST.
Pneumococcal infections remain a significant global cause of morbidity and death, particularly among children and the elderly. Rapid and dramatic changes in the epidemiology of this disease during the past decade in several developed countries followed the licensure and routine childhood administration of pneumococcal polysaccharide–protein conjugate vaccine (PCV). With PCV introduction in developing and middle-income countries, additional profound changes in pneumococcal ecology and disease epidemiology are likely. The disease burden and serotype distribution in the PCV era may be different than expected because of concomitant secular trends in pneumococcal disease, the impact of antibiotic use on pneumococcal strain ecology, and surveillance system attributes that can themselves affect analysis of epidemiologic features.
Not all pneumococcal serotypes are equally likely to cause disease; serotype distribution varies by age, disease syndrome, and geography. Geographic differences may be driven by variation in the burden of disease rather than by true serotype distribution differences. Most data on serotype distribution are related to pediatric invasive pneumococcal disease (IPD, defined as infection of a normally sterile site); much less information on global distribution is available on disease in adults. Among children <5 years of age, five to seven serotypes cause >60% of IPD cases in most parts of the world, seven serotypes (1, 5, 6A, 6B, 14, 19F, and 23F) account for ∼60% of cases in all areas of the world, but in any given region these seven serotypes may not all rank as the most common disease strains (Fig. 134-3). Some serotypes (e.g., types 1 and 5) not only tend to cause disease in areas with a high disease burden but also cause waves of disease in lower-burden areas (e.g., Europe) or outbreaks (e.g., in military barracks; meningitis in sub-Saharan Africa). The broader range of serotypes causing disease among adults than among children is apparent from a comparison of the coverage of existing multiserotype vaccines in different age groups. For example, data from the United States for 2006–2007 on the serotypes causing IPD indicated that a polysaccharide vaccine containing 23 serotypes (PPV23) would cover 84% of cases among children <5 years of age and 76% of those among persons 18–64 years of age but only 65% of those among persons ≥65 years of age.
Meta-analysis of available global serotype data, adjusted for regional disease incidence. The red line shows cumulative incidence, as indicated on the right-hand Y axis. (Source: Global Serotype Project Report for the Pneumococcal Advance Market Commitment Target Product Profile; available at http://www.vaccineamc.org/files/TPP_Codebook.pdf.)
Pneumococci are intermittent inhabitants of the healthy human nasopharynx and are transmitted by respiratory droplets. In children, pneumococcal nasopharyngeal ecology varies by geographic region, socioeconomic status, climate, degree of crowding, and particularly intensity of exposure to other children, with children in day-care settings having higher rates of colonization. In developed-world settings, children serve as the major vectors of pneumococcal transmission. By 1 year of age, ∼50% of children have had at least one episode of pneumococcal colonization. Cross-sectional prevalence data show rates of pneumococcal carriage ranging from 20% to 50% for children <5 years of age and from 5% to 15% for young and middle-aged adults; Fig. 134-4 shows relevant data from the United Kingdom. Data on colonization rates among healthy elderly individuals are limited. In developing-world settings, pneumococcal acquisition occurs much earlier, sometimes within the first few days after birth, and nearly all infants have had at least one episode of colonization by 2 months of age. Cross-sectional studies show that up to the age of 5 years, 70–90% of children carry S. pneumoniae in the nasopharynx, and a significant proportion of adults (sometimes >40%) are also colonized. Their high rates of colonization make adults an important source of transmission and may affect community transmission dynamics.
Prevalence of pneumococcal carriage in adults and children resident in the United Kingdom who had nasopharyngeal swabs collected monthly for 10 months (no seasonal trend; t test trend, >.05). (Data adapted from D Goldblatt et al, 2005.)
IPD develops when S. pneumoniae invades the bloodstream and seeds other organs or directly reaches the cerebrospinal fluid (CSF) by local extension. Pneumonia may follow aspiration of pneumococci, although only 10–30% of such cases are associated with a positive blood culture (and thus contribute to the measured burden of IPD). The dramatic variation of IPD rates with age is illustrated by data from the United States for 1998–1999, a period prior to PCV introduction. Rates of IPD were highest among children <2 years of age and among adults ≥65 years of age (188 and 60 cases per 100,000, respectively; Fig. 134-5). Since the introduction of PCV, IPD rates among infants and children in the United States have fallen by >75%, a decrease driven by the near elimination of vaccine-serotype IPD. A similar impact of PCV on vaccine-serotype IPD rates has been consistently observed in countries where PCV has been introduced into the routine pediatric vaccination schedule. However, changes in the non-vaccine-serotype IPD rate in various countries have been heterogeneous; the interpretation of this heterogeneity is a complex issue. In the United States, Canada, and Australia, rates of non-vaccine-serotype IPD have increased but the magnitude of the increase is generally small relative to the substantial reductions in vaccine-serotype IPD. In contrast, in other settings (e.g., Alaska Native communities and the United Kingdom), the reduction in vaccine-serotype IPD has been offset by notable increases in rates of disease caused by non-vaccine serotypes. Explanations for the heterogeneity of findings include replacement disease resulting from vaccine pressure, changes in clinical case investigation, secular trends unrelated to PCV use, antibiotic pressure selecting for resistant organisms, changes in surveillance or reporting systems, rapidity of introduction, and inclusion of a catch-up campaign. The roles and relative importance of these hypothesized mechanisms in driving the observed non-vaccine-serotype IPD trends and in explaining the observed heterogeneity among populations are not yet fully understood.
Rates of invasive pneumococcal disease before the introduction of pneumococcal conjugate vaccine, by age group: United States, 1998. [Source: CDC, Active Bacterial Core Surveillance/Emerging Infectious Program Network, 2000. Data adapted from MMWR 49(RR-9), 2000.]
Pneumonia is the most common of the serious pneumococcal disease syndromes and poses special challenges from a clinical and public health perspective. Most cases of pneumococcal pneumonia are not associated with bacteremia, and in these cases a definitive etiologic diagnosis is difficult. As a result, estimates of disease burden focus primarily on IPD rates and fail to include the major portion of the burden of serious pneumococcal disease. Among children, PCV trials designed to collect efficacy data on syndrome-based outcomes (e.g., radiographically confirmed pneumonia, clinically diagnosed pneumonia) have revealed the burden of culture-negative pneumococcal pneumonia.
The case-fatality ratios (CFRs) for pneumococcal pneumonia and IPD vary by age, underlying medical condition, and access to care. In addition, the CFR for pneumococcal pneumonia varies with the severity of disease at presentation (rather than according to whether or not the pneumonia episode is associated with bacteremia) and with the patient's age (from <5% among hospitalized patients 18–44 years old to >12% among those >65 years old, even when appropriate and timely management is available). Notably, the likelihood of death in the first 24 h of hospitalization did not change substantially with the introduction of antibiotics; this surprising observation highlights the fact that the pathophysiology of severe pneumococcal pneumonia among adults reflects a rapidly progressive cascade of events that often unfolds irrespective of antibiotic administration. Management in an intensive care unit can provide critical support for the patient through the acute period, with lower CFRs.
Rates of pneumococcal disease vary by season, with higher rates in colder than in warmer months in temperate climates; by gender, with males more often affected than females; and by risk group, with risk factors including underlying medical conditions, behavioral issues, and ethnic group. In the United States, some Native American populations (including Alaska natives) and African Americans have higher rates of disease than the general population; the increased risk is probably attributable to socioeconomic conditions and the prevalence of underlying risk factors for pneumococcal disease. Medical conditions that increase the risk of pneumococcal infection are listed in Table 134–1. Outbreaks of disease are well recognized in crowded settings with susceptible individuals, such as infant day-care facilities, military barracks, and nursing homes. Furthermore, there is a clear association between preceding viral respiratory disease (especially but not exclusively influenza) and risk of secondary pneumococcal infections. The significant role of pneumococcal pneumonia in the morbidity and mortality associated with seasonal and pandemic influenza is increasingly recognized.
Table 134–1 Clinical Risk Groups for Pneumococcal Infection |Favorite Table|Download (.pdf)
Table 134–1 Clinical Risk Groups for Pneumococcal Infection
|Clinical Risk Group||Examples|
|Asplenia or splenic dysfunction||Sickle cell disease, celiac disease|
|Chronic respiratory disease||Chronic obstructive pulmonary disease, bronchiectasis, cystic fibrosis, interstitial lung fibrosis, pneumoconiosis, bronchopulmonary dysplasia, aspiration risk, neuromuscular disease (e.g., cerebral palsy), severe asthma|
|Chronic heart disease||Ischemic heart disease, congenital heart disease, hypertension with cardiac complications, chronic heart failure|
|Chronic kidney disease||Nephrotic syndrome, chronic renal failure, renal transplantation|
|Chronic liver disease||Cirrhosis, biliary atresia, chronic hepatitis|
|Diabetes mellitus||Diabetes mellitus requiring insulin or oral hypoglycemic drugs|
|Immunocompromise/immunosuppression||HIV infection, common variable immunodeficiency, leukemia, lymphoma, Hodgkin's disease, multiple myeloma, generalized malignancy, chemotherapy, organ or bone marrow transplantation, systemic glucocorticoid treatment for >1 month at a dose equivalent to ≥20 mg/day (children, ≥1 mg/kg per day)|
|Cerebrospinal fluid leaks||…|
|Miscellaneous||Infancy and old age; prior hospitalization; alcoholism; malnutrition; cigarette smoking; day-care center attendance; residence in military training camps, prisons, homeless shelters|
Reduced pneumococcal susceptibility to penicillin was first noted in 1967, but not until the 1990s did reduced antibiotic susceptibility emerge as a significant clinical and public health issue, with an increasing prevalence of pneumococcal isolates resistant to single or multiple classes of antibiotics and a rising absolute magnitude of minimal inhibitory concentrations (MICs). Strains with reduced susceptibility to penicillin G, cefotaxime, ceftriaxone, macrolides, and other antibiotics are now found worldwide and account for a significant proportion of disease-causing strains in many locations, especially among children. Vancomycin resistance has not yet been observed in clinical pneumococcal strains. Lack of antimicrobial susceptibility is clearly related to a subset of serotypes, many of which disproportionately cause disease among children. The vicious cycle of antibiotic exposure, selection of resistant organisms in the nasopharynx, and transmission of these organisms within the community leading to difficult-to-treat infections and increased antibiotic exposure has been interrupted to some extent by the introduction and routine use of PCV. The clinical implications of pneumococcal antimicrobial nonsusceptibility are addressed below in the section on treatment.
Pneumococci colonize the human nasopharynx from an early age; colonization acquisition events are generally described as asymptomatic, but evidence exists to associate acquisition with mild respiratory symptoms, especially in the very young. From the nasopharynx, the bacteria spread either via the bloodstream to distant sites (e.g., brain, joint, bones, peritoneal cavity) or locally to mucosal surfaces where they can cause otitis media or pneumonia. Direct spread from the nasopharynx to the central nervous system (CNS) can occur in rare cases of skull base fracture, although most cases of pneumococcal meningitis are secondary to hematogenous spread. Pneumococci can cause disease in almost any organ or part of the body; however, otitis media, pneumonia, bacteremia, and meningitis are most common. Colonization is a relatively frequent event, yet disease is rare. In the nasopharynx, pneumococci survive in mucus secreted by epithelial cells, where they can avoid local immune factors such as leukocytes and complement. The mucus itself is a component of local defense mechanisms, and the flow of mucus (driven in part by cilia in what is known as the mucociliary escalator) effects mechanical clearance of pneumococci. While many colonization episodes are of short duration, longitudinal studies in adults and children have revealed persistent colonization with a specific serotype over many months. Colonization eventually results in the development of capsule-specific serum IgG, which is thought to play a role in mediating clearance of bacteria from the nasopharynx. IgG antibodies to surface exposed cell wall or secreted proteins also appear in the circulation in an age-dependent fashion or after colonization; the biological role of these antibodies is less clear. Recent acquisition of a new colonizing serotype is more likely to be associated with subsequent invasion, presumably as a result of the absence of type-specific immunity. Intercurrent viral infections make the host more susceptible to pneumococcal colonization, and pneumococcal disease in a colonized individual often follows perturbation of the nasopharyngeal mucosa by such infections. Local cytokine production after a viral infection is thought to upregulate adhesion factors in the respiratory epithelium, allowing pneumococci to adhere via a variety of surface adhesin molecules, including PsaA, PspA, CbpA, PspC, Hyl, pneumolysin, and the neuraminidases (Fig. 134-2). Adhesion coupled with inflammation induced by pneumococcal factors such as peptidoglycans and teichoic acids results in invasion. It is the inflammation induced by various bacterium-derived factors that is responsible for the pathology associated with pneumococcal infection. Cell wall–derived teichoic acids and peptidoglycans induce a variety of cytokines, including the proinflammatory cytokines interleukin (IL) 1, IL-6, and tumor necrosis factor (TNF), and activate complement via the alternative pathway. Polymorphonuclear leukocytes are thus attracted, and an intense inflammatory response is initiated. Pneumolysin is also important in local pathology, inducing proinflammatory cytokine production by local monocytes.
The pneumococcal capsule, consisting of polysaccharides with antiphagocytic properties due to resistance to the deposition of complement, plays an important role in pathogenesis. While most capsular types can cause human disease, certain capsular types are more commonly isolated from sites of infection. The reason for the dominance of some serotypes over others in IPD, as depicted in Fig. 134-3, is unclear.
As described above, intact respiratory epithelium and a host of nonspecific or innate immune factors (e.g., mucus, splenic function, complement, neutrophils, and macrophages) constitute the first line of defense against pneumococci. Physical factors such as the cough reflex and the mucociliary escalator are important in clearing bacteria from the lungs. Immunologic factors are critical as well: C-reactive protein (CRP) binds phosphorylcholine in the pneumococcal cell wall, inducing complement activation and leading to bacterial clearance; Toll-like receptor 2 (TLR2) recognizes both pneumococcal lipoteichoic acid and cell wall peptidoglycan; and in animal models, the absence of host TLR2 leads to more severe infection and impaired clearance of nasopharyngeal colonization. TLR4 appears to be necessary for the proinflammatory effect of pneumolysin on macrophages. The importance of TLR recognition is underlined by descriptions of an inherited deficiency of human IL-1 receptor–associated kinase 4 (IRAK-4) that manifests as an unusual susceptibility to infection with bacteria, including S. pneumoniae. IRAK-4 is essential for the normal functioning of several TLRs. Other factors that interfere with these nonspecific mechanisms (e.g., viral infections, cystic fibrosis, bronchiectasis, complement deficiency, and chronic obstructive pulmonary disease) all predispose to the development of pneumococcal pneumonia. Patients who lack a spleen or have abnormal splenic function (e.g., persons with sickle cell disease) are at high risk of developing overwhelming pneumococcal disease.
Acquired immunity induced via contact following colonization or through cross-reactive antigens rests largely on the development of serum IgG antibody specific for the pneumococcal capsular polysaccharide. Nearly all polysaccharides are T cell–independent antigens; B cells can make antibodies to such antigens without T cell help. However, in children <1–2 years old, such B cell responses are poorly developed. This delayed ontogeny of capsule-specific IgG in young children is associated with susceptibility to pneumococcal infection (Fig. 134-5). The extremely high risk of pneumococcal infection in the absence of serum immunoglobulin (i.e., in conditions such as agammaglobulinemia) highlights the important role of capsular antibody in protection against disease. Each serotype's capsule is chemically distinct; thus immunity tends to be serotype specific, although some cross-immunity exists. For example, conjugate vaccine–induced antibodies to serotype 6B prevent infection due to serotype 6A. However, cross-protection against serotypes within serogroups is not universal; for instance, antibodies to serotype 19F do not appear to confer protection against disease caused by serotype 19A. Antibodies to surface exposed or secreted pneumococcal proteins (such as pneumolysin, PsaA, and PspA) also appear in the circulation with increasing age of the host, but their functional significance remains unclear. Although data from murine models suggest that CD4+ T cells may play a role in preventing pneumococcal colonization and disease, these data have not yet been replicated in humans.
Approach to the Patient: Pneumococcal Infections
There is no pathognomonic presentation of pneumococcal disease; patients may present with a range of syndromes and with more than one clinical syndrome (e.g., pneumonia and meningitis). S. pneumoniae can infect nearly any body tissue, manifesting as disease ranging in severity from mild and self-limited to life-threatening. The differential diagnosis of common clinical syndromes such as pneumonia, otitis media, fever of unknown origin, and meningitis should always include pneumococcal infection. A microbiologically confirmed diagnosis is made in only a minority of pneumococcal cases since, in most circumstances (and especially in pneumonia and otitis media), fluid from the site of infection is not available for etiologic determination. Empirical therapy that includes appropriate treatment for S. pneumoniae is often indicated.
Algorithms for assessment and management of ill children have been developed for use in the developing world or in other settings where evaluation by a trained physician may not be feasible. Children who present with ominous signs such as inability to drink, convulsions, lethargy, and severe malnutrition are categorized as having very severe disease without further evaluation by the community health care worker, are given antibiotics, and are immediately referred to a hospital for diagnosis and management. Children who present with cough and tachypnea (the latter defined according to specific age strata) are further stratified into severity categories based on the presence or absence of lower chest wall indrawing and are managed accordingly with either antibiotics alone or antibiotics and referral to a hospital facility. Children with cough but no tachypnea are categorized as having a nonpneumonia respiratory illness.
The clinical manifestations of pneumococcal disease depend on the site of infection and the duration of illness. Clinical syndromes are classified as noninvasive (e.g., otitis media and nonbacteremic pneumonia) or invasive (e.g., bacteremic pneumonia). The pathogenesis of noninvasive illness involves contiguous spread from the nasopharynx or skin; invasive disease involves infection of a normally sterile body fluid or follows bacteremia.
Pneumonia is the most common serious pneumococcal syndrome and is considered invasive when associated with a positive blood culture. Pneumococcal pneumonia can present as a mild community-acquired infection at one extreme and as a life-threatening disease requiring intubation and intensive support at the other.
The presentation of pneumococcal pneumonia does not reliably distinguish it from pneumonia of other etiologies. In a subset of cases, pneumococcal pneumonia is recognized at the outset as associated with a viral upper respiratory infection and is characterized by the abrupt onset of cough and dyspnea accompanied by fever, shaking chills, and myalgias. The cough evolves from nonpurulent to productive of sputum that is purulent and sometimes tinged with blood. Patients may describe stabbing pleuritic chest pain and significant dyspnea indicating involvement of the parietal pleura. Among the elderly, the presenting clinical symptoms may be less specific, with confusion or malaise but without fever or cough. In such cases, a high index of suspicion is required because failure to treat pneumococcal pneumonia promptly in an elderly patient is likely to result in rapid evolution of the infection, with increased severity, morbidity, and risk of death.
Findings on Physical Examination
The clinical signs associated with pneumococcal pneumonia among adults include tachypnea (>30 breaths/min) and tachycardia, hypotension in severe cases, and fever in most cases (although not in all elderly patients). Respiratory signs are varied, including dullness to percussion in areas of the chest with significant consolidation, crackles on auscultation, reduced expansion of the chest in some cases as a result of splinting to reduce pain, bronchial breathing in a minority of cases, pleural rub in occasional cases, and cyanosis in cases with significant hypoxemia. Among infants with severe pneumonia, chest wall indrawing and nasal flaring are common. Nonrespiratory findings can include upper abdominal pain if the diaphragmatic pleura is involved as well as mental status changes, particularly confusion among elderly patients.
The differential diagnosis of pneumococcal pneumonia includes cardiac conditions such as myocardial infarction and heart failure with atypical pulmonary edema; pulmonary conditions such as atelectasis; and pneumonia caused by viral pathogens, mycoplasmas, Haemophilus influenzae, Klebsiella pneumoniae, Staphylococcus aureus, Legionella, or (in HIV-infected and otherwise immunocompromised hosts) Pneumocystis. In cases including abdominal symptoms, the differential diagnosis includes cholecystitis, appendicitis, perforated peptic ulcer disease, and subphrenic abscesses. The challenge in cases with abdominal symptoms is to remember to include pneumococcal pneumonia—a nonabdominal process—in the differential diagnosis.
Some authorities advocate treating uncomplicated, nonsevere, community-acquired pneumonia without determining the microbiologic etiology, given that this information is unlikely to alter clinical management. However, efforts to identify the cause of pneumonia are important when the disease is more severe and when the diagnosis of pneumonia is not clearly established. The gold standard for etiologic diagnosis of pneumococcal pneumonia is pathologic examination of lung tissue. In lieu of that procedure, evidence of an infiltrate on chest radiography warrants a diagnosis of pneumonia. However, cases of pneumonia without radiographic evidence do occur. An infiltrate can be absent either early in the course of the illness or with dehydration; upon rehydration, an infiltrate usually appears. The radiographic appearance of pneumococcal pneumonia is varied; it classically consists of lobar or segmental consolidation (Fig. 134-6) but in some cases is patchy. More than one lobe is involved in ∼30% of cases. Consolidation may be associated with a small pleural effusion or empyema in complicated cases. In children, “round pneumonia,” a distinctly spherical consolidation on chest radiography, is associated with a pneumococcal etiology. Round pneumonia is uncommon in adults. S. pneumoniae is not the only cause of such lesions; other causes, especially cancer, should be considered.
Chest radiograph depicting classic lobar pneumonia in the right lower lobe of an elderly patient's lung.
Blood drawn from patients with suspected pneumococcal pneumonia can be used for supportive or definitive diagnostic tests. Blood cultures are positive for pneumococci in a minority (<30%) of cases of pneumococcal pneumonia. Nonspecific findings include an elevated polymorphonuclear leukocyte count (>15,000/μL in most cases and upward of 40,000/μL in some), leukopenia in <10% of cases (a poor prognostic sign associated with a fatal outcome), and elevated values in liver function tests (e.g., both conjugated and unconjugated hyperbilirubinemia). Anemia, low serum albumin levels, hyponatremia, and elevated serum creatinine levels are all found in ∼20–30% of patients.
Urinary pneumococcal antigen assays have facilitated etiologic diagnosis. In adults, among whom the prevalence of pneumococcal nasopharyngeal colonization is relatively low, a positive pneumococcal urinary antigen test has a high predictive value. The same is not true for children, in whom a positive urinary antigen test can reflect the mere presence of S. pneumoniae in the nasopharynx.
Most cases of pneumococcal pneumonia are diagnosed by Gram's staining and culture of sputum. The utility of a sputum specimen is directly related to its quality and the patient's antibiotic treatment status.
Empyema is the most common focal complication of pneumococcal pneumonia, occurring in <5% of cases. When fluid in the pleural space is accompanied by fever and leukocytosis (even low-grade) after 4–5 days of appropriate antibiotic treatment for pneumococcal pneumonia, empyema should be considered. Parapneumonic effusions are more common than empyema, representing a self-limited inflammatory response to pneumonia. Pleural fluid with frank pus, bacteria (detected by microscopic examination), or a pH of ⩽7.1 indicates empyema and demands aggressive and complete drainage, usually through chest tube insertion.
Pneumococcal meningitis typically presents as a pyogenic condition that is clinically indistinguishable from meningitis of other bacterial etiologies. Meningitis can be the primary presenting pneumococcal syndrome or a complication of other conditions such as skull fracture, otitis media, bacteremia, or mastoiditis. Now that H. influenzae type b vaccine is routinely used, S. pneumoniae and Neisseria meningitidis are the most common bacterial causes of meningitis in both adults and children. Pyogenic meningitis, including that due to S. pneumoniae, is associated clinically with findings that include severe, generalized, gradual-onset headache, fever, and nausea as well as specific CNS manifestations such as stiff neck, photophobia, seizures, and confusion. Clinical signs include a toxic appearance, altered consciousness, bradycardia, and hypertension indicative of increased intracranial pressure. A small proportion of adult patients have Kernig's or Brudzinski's sign or cranial nerve palsies (particularly of the 3rd and 6th cranial nerves).
A definitive diagnosis of pneumococcal meningitis rests on the examination of CSF for (1) evidence of turbidity (visual inspection); (2) elevated protein level, elevated white blood cell count, and reduced glucose concentration (quantitative measurement); and (3) specific identification of the etiologic agent (culture, Gram's staining, antigen testing, or PCR). A blood culture positive for S. pneumoniae in conjunction with clinical manifestations of meningitis is also considered confirmatory. Among adults, detection of pneumococcal antigen in urine is considered highly specific because of the low prevalence of nasopharyngeal colonization in this age group.
The mortality rate for pneumococcal meningitis is ∼20%. In addition, up to 50% of survivors experience acute or chronic complications, including deafness, hydrocephalus, and mental retardation in children and diffuse brain swelling, subarachnoid bleeding, hydrocephalus, cerebrovascular complications, and hearing loss in adults.
S. pneumoniae can cause other invasive syndromes involving virtually any body site. These syndromes include primary bacteremia without other sites of infection (bacteremia without a source; occult bacteremia), osteomyelitis, septic arthritis, endocarditis, pericarditis, and peritonitis. The essential diagnostic approach is collection of fluid from the site of infection by sterile technique and examination by Gram's staining, culture, and—when relevant—capsular antigen assay or PCR. Hemolytic-uremic syndrome can complicate invasive pneumococcal disease.
The two major noninvasive syndromes caused by S. pneumoniae are sinusitis and otitis media; the latter is the most common pneumococcal syndrome and most often affects young children. The manifestations of otitis media include the acute onset of severe pain, fever, deafness, and tinnitus, most frequently in the setting of a recent upper respiratory tract infection. Clinical signs include a red, swollen, often bulging tympanic membrane with reduced movement on insufflation or tympanography. Redness of the tympanic membrane is not sufficient for the diagnosis of otitis media.
Pneumococcal sinusitis is also a complication of upper respiratory tract infections and presents with facial pain, congestion, fever, and—in many cases—persistent nighttime cough. A definitive diagnosis is made by aspiration and culture of sinus material; however, presumptive treatment is most commonly initiated after application of a strict set of clinical diagnostic criteria.
Treatment: Pneumococcal Infections
Historically, the activity of penicillin against pneumococci made parenteral penicillin G the drug of choice for disease caused by susceptible organisms, including community-acquired pneumonia. For susceptible strains, penicillin G remains the most commonly used agent, with daily doses ranging from 50,000 U/kg for minor infections to 300,000 U/kg for meningitis. Other parenteral β-lactam drugs, such as ampicillin, cefotaxime, ceftriaxone, and cefuroxime, can be used against penicillin-susceptible strains but offer little advantage over penicillin. Macrolides and cephalosporins are alternatives for penicillin-allergic patients. While agents such as clindamycin, tetracycline, and trimethoprim-sulfamethoxazole exhibit some activity against pneumococci, resistance to these agents is frequently encountered in different parts of the world.
Penicillin-resistant pneumococci were first described in the mid-1960s, at which point tetracycline- and macrolide-resistant strains had already been reported. Multidrug-resistant strains were first described in the 1970s, but it was during the 1990s that pneumococcal drug resistance reached pandemicproportions. The use of antibiotics selects for resistant pneumococci, and strains resistant to β-lactam agents and to multiple drugs are now found all over the world. The emergence of high rates of macrolide and fluoroquinolone resistance has also been described.
The molecular basis of penicillin resistance in S. pneumoniae is the alteration of penicillin-binding protein (PBP) genes by transformation and horizontal transfer of DNA from related streptococcal species. Such alteration of PBPs results in lower affinity for penicillins. Depending on the specific PBP(s) and the number of PBPs altered, the level of resistance ranges from intermediate to high. For many years, penicillin susceptibility breakpoints have been defined by MICs as follows: susceptible, ⩽0.06 μg/mL; intermediate, 0.12–1.0 μg/mL; and resistant, ⩾2.0 μg/mL. However, in vitro results often were not predictive of the response of a patient to treatment for pneumococcal diseases other than meningitis. New recommendations have been based on the revised penicillin G breakpoints established in 2008 by the Clinical and Laboratory Standards Institute. For IV treatment of meningitis with at least 24 million units per day in 8 divided doses, the susceptibility breakpoint remains ⩽0.06 μg/mL, and MICs of ⩾0.12 μg/mL indicate resistance. For IV treatment of nonmeningeal infections with 12 million units per day in 6 divided doses, the breakpoints are ⩽2 μg/mL for susceptible organisms, 4 μg/mL for intermediate organisms, and ⩾ 8 μg/mL for resistant organisms; a dosage of 18–24 million units per day is recommended for strains with MICs in the intermediate category. The original breakpoints remain the same for oral treatment of nonmeningeal infections with penicillin V.
Although guidelines for antibiotic therapy should be driven in part by local patterns of resistance, guidelines from national organizations in many countries (e.g., the Infectious Diseases Society of America/American Thoracic Society, the British Thoracic Society, and the European Respiratory Society) lay out evidence-based approaches. The following guidelines for individual sepsis syndromes are based on those advocated by the American Academy of Pediatrics and published in the 2009 Red Book.
Meningitis Likely or Proven to Be Due to S. Pneumoniae
As a result of the increased prevalence of resistant pneumococci, first-line therapy for persons ≥1 month of age is a combination of vancomycin (adults, 30–60 mg/kg per day; infants and children, 60 mg/kg per day) and cefotaxime (adults, 8–12 g/d in 4–6 divided doses; children, 225–300 mg/kg per day in 1 dose or 2 divided doses) or ceftriaxone (adults, 4 g/d in 1 dose or 2 divided doses; children, 100 mg/kg per day in 1 dose or 2 divided doses). If children are hypersensitive to β-lactam agents (penicillins and cephalosporins), rifampin (adults, 600 mg/d; children, 20 mg/d in 1 dose or 2 divided doses) can be substituted for cefotaxime or ceftriaxone. A lumbar puncture should be considered after 48 h if the organism is not susceptible to penicillin and information on cephalosporin sensitivity is not yet available, if the patient's clinical condition does not improve or deteriorates, or if dexamethasone has been administered and may be compromising clinical evaluation. When antibiotic sensitivity data become available, treatment should be modified accordingly. If the isolate is sensitive to penicillin, vancomycin can be discontinued and penicillin can replace the cephalosporin, or cefotaxime or ceftriaxone can be continued alone. If the isolate displays any resistance to penicillin but is susceptible to the cephalosporins, vancomycin can be discontinued and cefotaxime or ceftriaxone continued. If the isolate exhibits any resistance to penicillin and is not susceptible to cefotaxime and ceftriaxone, vancomycin and high-dose cefotaxime or ceftriaxone can be continued; rifampin may be added as well if the isolate is susceptible and the patient's clinical condition is worsening, if the CSF remains positive for bacteria, or if the MIC of the cephalosporin in question against the infecting strain is high. Some physicians advocate the use of glucocorticoids in children >6 months old, but this recommendation remains controversial and is not universally considered the standard of care. Glucocorticoids significantly reduce rates of mortality, severe hearing loss, and neurologic sequelae in adults and should be administered to those with community-acquired bacterial meningitis. If dexamethasone is given to either adults or children, it should be administered before or in conjunction with the first antibiotic dose.
Invasive Infections (Excluding Meningitis)
In previously well children with noncritical illness, antibiotic therapy with a recommended antibiotic should be instigated at the following dosages: penicillin G, 250,000–400,000 units/kg per day (in divided doses 4–6 h apart); cefotaxime, 75–100 mg/d (doses 8 h apart); or ceftriaxone, 50–75 mg/d (doses 12–24 h apart). For critically ill children, including those with myocarditis or multilobular pneumonia with hypoxia or hypotension, vancomycin may be added if the isolate may possibly be resistant to β-lactam drugs, with its use reviewed once susceptibility data become available. If the organism is resistant to β-lactam agents, therapy should be modified on the basis of clinical response and susceptibility to other antibiotics. Clindamycin or vancomycin can be used as a first-line agent for children with severe β-lactam hypersensitivity, but vancomycin should not be continued if the organism is shown to be sensitive to other non-β-lactam antibiotics.
For outpatient management, amoxicillin (1 g every 8 h) provides effective treatment for virtually all cases of pneumococcal pneumonia. Neither cephalosporins nor quinolones, which are far more expensive, offer any advantage over amoxicillin. Levofloxacin (500–750 mg/d as a single dose) and moxifloxacin (400 mg/d as a single dose) are also highly likely to be effective in the United States except in patients who come from closed populations where these drugs are used widely or who have themselves been treated recently with a quinolone. Clindamycin (600–1200 mg/d every 6 h) is effective in 90% of cases and azithromycin (500 mg on day 1 followed by 250–500 mg/d) or clarithromycin (500–750 mg/d as a single dose) in 80% of cases. Treatment failure resulting in bacteremic disease due to macrolide-resistant isolates has been amply documented in patients given azithromycin empirically. As noted above, rates of resistance to all these antibiotics are relatively low in some countries and much higher in others; high-dose amoxicillin remains the best option worldwide.
The optimal duration of treatment for pneumococcal pneumonia is uncertain, but its continuation for at least 5 days once the patient becomes afebrile appears to be a prudent approach. Cases with a second focus of infection (e.g., empyema or septic arthritis) require longer therapy.
Amoxicillin (80–90 mg/kg per day) is recommended for children with acute otitis media except in situations where observation and symptom-based treatment without antibiotics are advocated. These situations include nonsevere illness and an uncertain diagnosis in children 6 months to 2 years of age and nonsevere illness (even if the diagnosis seems certain) in children >2 years of age. Although the optimal duration of therapy has not been conclusively established, a 10-day course is recommended for younger children and for children with severe disease at any age. For children >6 years old who have mild or moderate disease, a course of 5–7 days is considered adequate. Patients whose illness fails to respond should be reassessed at 48–72 h. If acute otitis media is confirmed and antibiotic treatment has not been started, administration of amoxicillin should be commenced. If antibiotic therapy fails, a change is indicated. Failure to respond to second-line antibiotics as well indicates that myringotomy or tympanocentesis may need to be undertaken in order to obtain samples for culture.
The above recommendations can also be followed for the treatment of sinusitis. Detailed information on the further management of these conditions in children has been published by the American Academy of Pediatrics and the American Academy of Family Physicians.
Measures to prevent pneumococcal disease include vaccination against S. pneumoniae and influenza viruses, reduction of comorbidities that increase the risk of pneumococcal disease, and prevention of antibiotic overuse, which fuels pneumococcal resistance.
Capsular Polysaccharide Vaccines
The 23-valent pneumococcal polysaccharide vaccine (PPV23), containing 25 μg of each capsular polysaccharide, has been licensed for use since 1983. Recommendations for its use vary by country. The U.S. Advisory Committee on Immunization Practices recommends PPV23 for all persons ≥65 years of age and for those 2–64 years of age who have underlying medical conditions that put them at increased risk for pneumococcal disease or severity (Table 134–1; see also http://www.cdc.gov/vaccines/recs/schedules/default.htm). Revaccination 5 years after the first dose is recommended for persons >2 years of age who have underlying medical conditions but not routinely for those whose only indication is an age of ≥65 years. PPV23 does not induce an anamnestic response, and antibody concentrations wane over time; thus revaccination is particularly important for individuals with conditions resulting in loss of antibody. Concerns about repeated revaccination have focused on safety (i.e., local reactions) and the induction of immune hyporesponsiveness. Neither the clinical relevance nor the biological basis of hyporesponsiveness is clear, but, given the possibility of its occurrence, more than one revaccination has not been recommended.
The effectiveness of PPV23 against IPD, pneumococcal pneumonia, all-cause pneumonia, and death is controversial, with wide variation in observations. The many published meta-analyses of PPV efficacy have often reached opposing conclusions with regard to a given clinical entity. Generally, observational studies cite greater effectiveness than do controlled clinical trials. The consensus is that PPV is effective against IPD but is less effective or ineffective against nonbacteremic pneumococcal pneumonia. However, published trials, observational studies, and meta-analyses contradict this view. Efficacy is often lower in the elderly and in immunodeficient patients whose condition is associated with reduced antibody responses to vaccines than in younger, healthier populations. When PPV is effective, the duration of protection following a single dose of vaccine is estimated to be ∼5 years.
What is not disputed is that improved pneumococcal vaccines are needed for adults. Even in the setting of routine vaccination of infants (which indirectly protects adults from vaccine-serotype strains), disease caused by serotypes not represented in the vaccine continues to be responsible for a significant burden of disease among adults.
Polysaccharide–Protein Conjugate Vaccines
Infants and young children respond poorly to PPV, which contains T cell–independent antigens. Consequently, another class of pneumococcal vaccines, the PCVs, were developed specifically for infants and young children. The first product, a 7-valent PCV, was licensed in 2000 in the United States. As of 2010, three PCV products—containing 7, 10, and 13 serotypes, respectively—were commercially available. The serotypes included in these PCV formulations are important causes of IPD and antibiotic resistance among young children. Randomized controlled trials have demonstrated a high degree of efficacy of PCVs against vaccine-serotype IPD as well as efficacy against pneumonia, otitis media, nasopharyngeal colonization, and all-cause mortality. PCVs are recommended by the World Health Organization for inclusion in routine childhood immunization schedules worldwide, especially in countries with high infant mortality rates.
The United States was the first country to introduce PCV and therefore has the longest experience with its community-wide effects. The introduction of PCV in the United States has resulted in a >90% reduction in vaccine-serotype IPD among the whole population (Fig. 134-7). This decline has been noted not only in those age groups immunized but also in adults and is attributable to the near elimination of vaccine-serotype nasopharyngeal colonization in immunized infants, which reduces spread to adults. This protection of unimmunized community members through vaccination of a subset of the community is termed the indirect effect. Increases in colonization with—and concomitantly in disease due to—non-vaccine-serotype strains (i.e., replacement colonization and disease) have been seen; however, the absolute rate increases in IPD caused by non-vaccine serotypes are generally small, especially relative to decreases in vaccine-serotype IPD (see “Epidemiology,” above). Since vaccine-serotype strains are more commonly resistant to antibiotics than are non-vaccine serotypes, use of PCV has also resulted in dramatic declines in the proportion and absolute rates of drug-resistant pneumococcal disease. The recommendations of the Advisory Committee on Immunization Practices for the use of conjugate vaccines can be found at http://www.cdc.gov/MMWR/pdf/wk/mm5909.pdf. Recently, PCV has been shown to prevent pneumococcal infection in HIV-infected adults.
Changes in invasive pneumococcal disease (IPD) incidence, by serotype group, among children <5 years old (top) and adults >65 years old (bottom), 1998–2007. *7-Valent pneumococcal conjugate vaccine (PCV7) was introduced in the United States for routine administration to infants and young children during the second half of 2000. (Reprinted with permission from Pilishvili et al, 2010.)
Other Prevention Strategies
Pneumococcal disease can also be averted through the prevention of illnesses that predispose individuals to pneumococcal infections. Relevant measures include influenza vaccination and improved management and control of diabetes, HIV infection, heart disease, and lung disease. Finally, the reduction of antibiotic misuse is a strategy for the prevention of pneumococcal disease in that antimicrobial resistance directly and indirectly perpetuates organism transmission and disease in the community.