Current Diagnosis & Treatment in Infectious Diseases

Section V - Bacterial Infections

47. Streptococcus pneumoniae

Jeffrey N. Martin MD, MPH

Essentials of Diagnosis

  • Most common infections include pneumonia, meningitis, sinusitis, and otitis media.
  • Predisposing factors include extremes of age (ie, infants and elderly) and underlying host defects in antibody, complement, and splenic function.
  • Transmission is human to human both in the community and nosocomially.
  • Pneumonia: productive cough, fever, chills, sweats, and dyspnea; lobar or segmental consolidation on chest x-ray; lancet-shaped gram-positive diplococci on Gram stain of sputum or growth of S pneumoniaefrom sputum, blood, or pleural fluid.
  • Meningitis: headache, stiff neck, fever, chill, and photophobia; nuchal rigidity, Kernig's or Brudzinski's sign; lancet-shaped gram-positive diplococci on Gram stain of cerebrospinal fluid (CSF), growth of S pneumoniaefrom CSF or blood, or positive counter immunoelectrophoresis or latex agglutination of CSF.

General Considerations

  • Epidemiology.S pneumoniae may exist in humans as either a nasopharyngeal colonist or as a pathogen in one of many clinical syndromes (Box 47-1). Although S pneumoniae has been found in other mammals, humans are thought to be the principal reservoir. As a colonist, S pneumoniae is found in up to 40% of children and 10% of adults. Infants typically acquire their first colonizing isolate at 6 months. Intermittent colonization then occurs throughout childhood and adult years. There is a seasonal pattern to colonization, with the highest prevalence occurring in winter. Pneumococci are transmitted from human to human by extensive direct contact with secretions harboring the organism. Persons residing in environments that promote such extensive close contact, such as day care centers, prisons, homeless shelters, and military barracks, have historically been at risk for pneumococcal infection.

In the United States, S pneumoniae is estimated to account for 500,000 cases of pneumonia, 50,000 cases of bacteremia, 3000 cases of meningitis, and 7 million cases of acute otitis media. Of these syndromes, bacteremia (with or without a site of primary infection such as pneumonia or meningitis) has the most clearly elucidated epidemiology because of its precise blood culture-based definition. The incidence of bacteremia is strongly age related. A surveillance study in South Carolina estimated the incidence of bacteremia among infants, young adults, and elderly (age ≥ 70) to be 160, 5, and 70 cases per 100,000 persons, respectively. African-Americans, American Indians, and Alaskan Natives are at highest risk for bacteremia. As is true for S pneumoniae colonization, there is a distinct seasonality to bacteremia with peaks coming at midwinter. Overall, pneumococcal disease accounts for ~ 40,000 annual deaths in the United States. The case fatality rate for bacteremia is 15–20%.

In addition to the numerical incidence of pneumococcal disease, it is instructive to look at the prominent role S pneumoniae plays in several common syndromes. In a variety of studies that prospectively identify causes of community-acquired pneumonia, S pneumoniae is routinely the most common detectable agent. For example, in patients who require hospitalization for pneumonia, S pneumoniae accounts for up to one-third of cases. Aside from areas of the world experiencing meningococcal epidemics, S pneumoniae is the most common organism of meningitis in adults and now, with the advent of Haemophilus influenzae type b immunization, is also most common in children. One-third to one-half of all cases of acute otitis media with an identifiable etiology is caused by S pneumoniae.

Any of the syndromes caused by S pneumoniae can occur in normal hosts, but frequently one or more predisposing conditions exist. The best characterized of these are underlying host immunologic defects in antibody, complement, and splenic function. Because of the importance of antibodies in the opsonization of the encapsulated pneumococcus, individuals with defective antibody function (eg, congenital agammaglobulinemia, common variable hypogammaglobulinemia, selective immunoglobulin-G subclass deficiency, multiple myeloma, chronic lymphocytic leukemia, or lymphoma) are at significant risk for invasive infections. Various early complement deficiencies as well as either congenital or acquired asplenia (eg, iatrogenic or sickle cell disease) also convey substantial risk. Asplenic patients deserve special note because of the rapidity at which they can clinically deteriorate from unchecked bacteremia.

In contemporary urban settings, HIV infection is responsible for a large burden of pneumococcal disease. HIV disease, through a variety of host defense defects, places infected persons at ~ 200-fold-greater risk for invasive pneumococcal infection. HIV-infected patients who are African-American, who have a CD4 lymphocyte count < 200 × 106/liter, or who have a history of pneumonia have particularly elevated risk. The presence of pneumococcal bacteremia in patients with no clinically apparent host immunodeficiency should prompt the clinician to obtain patient consent for HIV testing.

Various other chronic illnesses, such as alcoholism, cirrhosis, renal insufficiency, nephrotic syndrome, chronic pulmonary disease, congestive heart failure, and diabetes, place patients at significant risk for pneumococcal disease. Likewise, patients who have received organ or bone marrow transplants or who are being treated with alkylating agents, antimetabolites, or glucocorticoids are also at risk. Finally, a particular anatomic defect, basilar skull fracture, carries risk for pneumococcal meningitis.

  • Microbiology.S pneumoniae is a gram-positive coccus, but typically the organism grows as pairs in what is termed a lancet-shaped diplococcus. In truth, this shape represents two lancets joined on end, pointing away from each other. In liquid medium, pneumococci may also be found growing in chains. On blood agar plates, pneumococcal colonies are often umbilicated (ie, have central dimpling) and are sometimes grossly mucoid. They are surrounded by a greenish zone, known as α-hemolysis, which is the result of hemoglobin breakdown by pneumococcus-produced α-hemolysin. In the clinical microbiology laboratory, α-hemolysis is the first clue in distinguishing S pneumoniae from a variety of other species in a mixed flora sample. In addition to α-hemolysis, the microbiology laboratory uses catalase negativity, solubility in bile salts, and susceptibility to optochin to identify S pneumoniae (Table 47-1).

S pneumoniae is easy to identify by the above criteria when it grows in normally sterile body sites (eg, blood or pleural fluid). Detection in a mixed-flora environment such as sputum, however, requires a skilled and persevering clinical microbiologist. Given the abundance of other α-hemolytic oral species, it is incumbent on the microbiologist to look carefully for the characteristic umbilicated or mucoid appearance of S pneumoniae and to subculture a number of α-hemolytic colonies. The clinician can be motivating in this endeavor by communicating the patient's clinical presentation to the microbiologist and by carefully reviewing the Gram stain of the original specimen for the presence of lancet-shaped diplococci.

Once called Diplococcus pneumoniae, S pneumoniae is now known to belong to the Streptococcus genus. Like most other members of this genus, S pneumoniae is non-spore forming and nonmotile. Unlike many streptococcal species, S pneumoniae does not have a Lancefield serogroup. Yet, within the S pneumoniae species, there are 90 identified serotypes. Two numbering schemes for serotyping exist. In the American system, serotypes are numbered starting from 1 to 90 in the order in which they were first described. In the more widely used Danish system, serotypes are grouped based on antigenic similarities and given both a number and a letter (eg, 19A, 19B, and 19C). The basis of the serotype is antigenic differences in the polysaccharides that constitute the pneumococcal external capsule. Almost all clinical strains, with the uncommon exception of some isolates from conjunctivitis, have a capsule. In the microbiology laboratory, the interaction of serotype-specific antibody with capsule results in the enhancement of the microscopic appearance of the capsule, which is known as the quellung reaction. Although serotyping is no longer an important clinical tool, it remains useful for epidemiologic studies; surveys of the most prevalent serotypes found in invasive infections have informed the composition of the current polyvalent vaccine.

Table 47-1. Clues to diagnosis of Streptococcus pneumoniae infection

o   Gram–positive coccus; grows as lancet-shaped diplo-cocci or in chains

o   Grows readily in conventional blood culture or solid media, but fastidious nature necessitates rapid transport of clinical material to microbiology laboratory for processing

o   α–Hemolytic (greenish) on solid media; colonies are often umbilicated or mucoid

o   Catalase-negative

o   Bile salt soluble

o   Susceptible to optochin

o   Counter immunoelectrophoresis or latex agglutination may be useful in evaluation of cerebrospinal fluid that was obtained after antibiotics were administered

Whereas the external capsule is a distinguishing feature of S pneumoniae, its cell wall is similar to that of other streptococci. Multilayered peptidoglycan, a heteropolymer of repeating N-acetylglucosamine and N-acetyl muramic acid, is the principal component. Peptides, attached to N-acetyl muramic acid, are cross-linked by trans- and carboxypeptidases to provide structure to the cell wall. These peptidases are also known as penicillin-binding proteins because they are the binding sites for a variety of β-lactam antibiotics. The covalent binding of β-lactams inactivates these enzymes resulting in the eventual demise of the organism. Teichoic acid (ribotol- or glycerol-phosphate polymers) is the other principal component of the cell wall and is often covalently bound to peptidoglycan. Lipoteichoic acid exists as a component of both the cell membrane and cell wall. There are other minor components of the cell wall, some of which are common to all streptococci, and some of which, like C-substance (a polysaccharide), are unique to S pneumoniae. Various soluble products are produced by S pneumoniae and are discussed in the next section.

  • Pathogenesis.Diseases caused by S pneumoniae are principally the result of an exuberant inflammatory response that the host mounts to the organism. The critical factor in the success of the pneumococcus is its ability to evoke this response but simultaneously, at least until specific anticapsular antibody is produced, to evade nonspecific attempts at phagocytosis. Highly destructive toxins, analogous to those produced by Streptococcus pyogenes or gram-negative species, are not present although some toxins are produced and play a secondary role. Although details vary slightly depending on the organ system involved, the pathogenic progression consists of nasopharyngeal adherence and colonization, spread to adjacent organs with local replication, evasion of nonspecific clearance mechanisms, and finally activation of the host inflammatory reaction.

The initiating event in all cases is nasopharyngeal colonization, in which bacterial surface adhesins join to epithelial cell receptors that contain the disaccharide GlcNAcβ1-4Gal. After pneumococci are established in the nasopharynx, they may then gain entry into and replicate in contiguous structures such as the sinuses, eustachian tubes, or bronchi. They may also penetrate the nasopharyngeal mucosa, even without a clinically demonstrable focus of infection, and achieve access to the systemic circulation via the cervical lymphatics. Entry into the lungs is limited in normal hosts by a functioning glottis and larynx, but entry is enhanced when these mechanisms are disturbed, as seen with alcohol intoxication, convulsions, anesthesia, or stroke.

Even after pneumococci gain entry to the above structures, infection rarely occurs because they are typically cleared in the normal host by nonspecific mechanisms (eg, mucociliary motion or cough reflex). Replication can proceed, however, when normal clearance is impaired, as seen, for example, when eustachian tubes or sinus orifices are congested (eg, from viral infection or allergy) or bronchial clearance is altered (eg, from chronic effects of smoking or acute effects of viral infection). In the lung, various mechanisms are now being described that may enhance pneumococcal adherence and replication. These include the finding that pneumococci bind to immobilized fibronectin that is exposed during tissue injury and that cytokines may induce pneumocytes to express the receptor for platelet-activating factor—a receptor that also binds pneumococcal C-substance.

Once established in its target organs, S pneumoniae is able to replicate unimpeded because of its ability to evade nonspecific phagocytosis by polymorphonuclear cells. The external capsule provides this crucial protection. As such, the capsule is an essential factor in the virulence of the organism. The exact mechanism of this protective effect is not known; it is likely a combination of the lack of capsular receptors on phagocytes, repellent electrochemical forces on the capsule, and the ability of the capsule to mask opsonizers such as antibody and complement. Whatever the exact mechanism, S pneumoniae organisms that are genetically engineered to lack capsules are avirulent.

Once it begins replicating, S pneumoniae causes disease by evoking an intense inflammatory reaction. Both the peptidoglycan and teichoic acid components of the cell wall and capsular polysaccharide can activate the alternative pathway of complement. Nonspecific antibodies to cell wall polysaccharides (formed to ubiquitous streptococcal species) are able to activate complement by the classical pathway. Together, these means of activating complement result in the attraction of numerous leukocytes and abundance of exudative fluid. In the lungs, this accumulation of pneumococci and inflammatory material leads to consolidation of alveoli, radiolucency, and impaired gas exchange—the defining characteristics of pneumonia. Although abscess formation is rare, the expanding volume of infected material can spread to uninvolved areas via Kohn's pores. Extensive direct spread can result in empyema or pericardial infection. If the infection is not contained, pneumococci can spread via lymphatics to hilar lymph nodes, the thoracic duct, and finally into the systemic circulation where metastatic infection may occur in the meninges, peritoneum, joints, or endocardium.

As mentioned, S pneumoniae may gain access to the subarachnoid space via hematogenous spread and choroid seeding from either a nasopharyngeal or pulmonary focus. Less commonly, meninges may become infected via direct extension from an infected sinus or middle ear. As in the lungs, an intense inflammatory response is evoked by the presence of pneumococci in the subarachnoid space. Interleukin-1 and tumor necrosis factor, released by macrophages responding to the insult, play a prominent role and contribute to increased blood-brain barrier permeability. The end result is mounting intracranial pressure and ultimately diminished blood flow to the brain.

Before the advent of antibiotics, this pathogenic process could be arrested by the development of specific anticapsular antibody. This typically occurs at 5–8 days after the onset of infection. Hence, outcome depended on whether antibody production occurred before the patient succumbed to the severe manifestations of infection. In the contemporary era, it is now appreciated that the pathogenic process can also be arrested by the administration of antibiotics. However, the persistently high case fatality rate associated with pneumococcal disease (especially bacteremia) despite antibiotic administration bespeaks the continued contributory role of host defenses in successfully responding to infection.

S pneumoniae produces several toxins that are currently thought to play a secondary role in pathogenesis. Pneumolysin is a toxin with two functions. It can insert into cell membranes of polymorphonuclear cells and ciliated epithelium, thereby inhibiting their function, and it can activate the classical complement pathway. Other products such as pneumococcal surface protein, hemolysin, and autolysin are also contributory because mutants lacking these typically are less virulent. Autolysin in particular is important because autolysis of organisms may enhance the release and expression of a variety of proinflammatory factors (eg, cell wall components) that were discussed above.

BOX 47-1 Syndromes Caused by Streptococcus pneumoniae

More Common

· Pneumonia (with or without bacteremia)

· Acute sinusitis

· Otitis media1

· Meningitis

· Primary bacteremia1

Less Common

· Empyema

· Septic arthritis

· Peritonitis

· Endocarditis

· Pericarditis

· Cellulitis

· Endometritis

· Osteomyelitis

· Brain abscess

1More common in children than adults.

BOX 47-2 Clinical Findings in Pneumococcal Pneumonia




Laboratory Tests1

More Common

· Fever

· Chills

· Sweats

· Cough

· Rusty sputum production

· Dyspnea

· Fever, tachycardia, tachypnea

· Consolidative findings on chest exam

· Hypoxemia

· Elevated leukocyte count

Less Common2

· Pleuritic chest pain

· Nausea, vomiting, diarrhea

· Confusion

· Hypothermia

· Hypotension

· Cyanosis

· Oral herpes lesions

· Nuchal rigidity

· Abdominal distension and absent bowel sounds

· Pleural rub

· Murmur

· Altered mental status

· Low serum bicarbonate, anion gap

· Elevated total bilirubin and lactate

1Radiographic and microbiologic findings are discussed in the text.
2Many of these findings represent the presence of concomitant pneumococcal disease in other organ systems.



This section focuses on the clinical findings, diagnosis, and treatment of pneumococcal pneumonia. Please see Chapters 9 and 10 for a general discussion of pneumonia and upper respiratory infection.

Clinical Findings

  • Signs and Symptoms.The so-called classic presentation of pneumococcal pneumonia consists of an abrupt onset of a shaking chill followed shortly after by fever and production of rusty sputum (Box 47-2). This classic presentation, however, is not the most common. Typically, patients present because of a change in preexisting symptoms. Many have had several days of a viral upper respiratory infection marked by coryza, sore throat, and nonproductive cough, or they have chronic pulmonary disease with its attendant persistent productive cough. In these patients, the development of pneumonia is heralded by the onset of fever (which can be as high as 40°C), chills, sweats, malaise, dyspnea, and a change in the nature of the cough. Cough may either turn productive in those without prior production or increase in volume and change color in those with prior production. Less common symptoms include pleuritic chest pain and—for reasons that are poorly understood—nausea, vomiting, and diarrhea. Some patients, particularly the elderly, will present with merely an alteration in mental status. Thus, the very common entity of pneumococcal pneumonia must be entertained for a wide variety of clinical presentations.

On physical examination, persons with pneumococcal pneumonia are usually markedly ill appearing and may be cyanotic. Altered vital signs such as elevated temperature, tachycardia, and tachypnea (> 22 breaths/min) are common but, as noted above, can be absent in the elderly. In fact, in the elderly, hypothermia may be present. With the exception of the appearance of oral herpes lesions, the majority of the physical findings relate to the lungs. Dullness to percussion, increased tactile fremitus, bronchophony, whispered pectoriloquy, and egophony are sometimes present as clues to underlying consolidation, but often only rales are heard. It must be remembered, however, that even the presence of rales in an otherwise normal individual must be considered an abnormal finding and prompt further radiographic consideration of pneumonia. Conditions associated with pneumonia, meningitis (eg, nuchal rigidity), and endocarditis (eg, murmur) should also be routinely evaluated.

  • Laboratory Findings.An elevated leukocyte count, often with early forms (left-shifted), is common. However, in certain patient populations (eg, alcohol abusers and persons with HIV infection), the absence of a high leukocyte count does not have sufficient negative predictive value to exclude pneumococcal disease. In acutely ill patients, a low serum bicarbonate concentration and an anion gap may be seen. This should prompt arterial blood gas determination to document the presence and extent of acidemia. Although determination is not necessarily indicated in all patients presenting with suspected pneumonia, an elevated bilirubin level and lactate dehydrogenase are prevalent. Pulse oximetry on room air should be performed in all patients. Individuals with low pulse oximetry readings and for whom acid-base status is unsettled should have arterial blood gas measurements.

Two sets of blood cultures should be performed on all patients with suspected pneumonia in that this may be the only way (aside from pleural fluid culture) to diagnose definitively the etiologic agent. Nevertheless, the majority of patients with pneumococcal pneumonia are not bacteremic. Hence, particular attention must be paid to the adequate collection and processing of expectorated sputum. Although S pneumoniae can be a colonist, its growth in sputum culture in a patient with clinically diagnosed pneumonia is usually satisfactory proof of pneumococcal pneumonia. The sensitivity of culture, however, is often solely dependent on the rigor practiced by microbiology staff in the identification and subculturing of α-hemolytic colonies. Findings on Gram stain of the sputum can also be helpful but require careful interpretation. The presence of characteristic gram-positive lancet-shaped diplococci accompanied by an abundance of leukocytes (> 25 per low [100×]-power field) and a paucity of epithelial cells (< 10 per low-power field) on Gram stain suggests pneumococcal pneumonia, but the absence of this Gram stain pattern does not exclude pneumococcal disease. Although there are increasing regulations regarding clinician-prepared Gram stains, clinicians should routinely review stains that are prepared by microbiology staff.

  • Imaging.The sine qua non of pneumococcal pneumonia is the presence of an infiltrate on chest radiography. The clinical diagnosis of pneumonia has to be seriously questioned if, after 24 h of observation and volume repletion, a patient has not yet developed an infiltrate. A variety of radiographic findings can be observed. Lobar or segmental consolidation is classic but is less common than subsegmental involvement. Multiple subsegmental consolidation may be seen, and, occasionally, multilobar consolidation may be present. Pleural effusion is the other common finding, and the failure of a patient to improve necessitates prompt investigation of this fluid for the presence of empyema. Lung abscesses are rare, and their presence suggests serotype-3 infection. More commonly, the presence of an abscess suggests preexisting cavitary disease or coinfection with other abscess-producing organisms.
  • Differential Diagnosis.In patients presenting with community-acquired pneumonia (ie, acute pulmonary infection in a patient who has not been hospitalized nor has resided in a long-term-care facility in the 14 days prior to the onset of symptoms), a number of agents must be considered along with S pneumoniae in the differential diagnosis. H influenzae, Branhamella (Moraxella) catarrhalis, Chlamydia pneumoniae, and Mycoplasma pneumoniae are the most common agents, but, in certain populations, Legionella spp., Klebsiella pneumoniae, Staphylococcus aureus, Escherichia coli, Chlamydia psittaci, Coxiella burnetti, Mycobacterium tuberculosis, Coccidioides immitis, Histoplasma capsulatum, Blastomyces dermatitidis, Cryptococcus neoformans, and several respiratory viruses (eg, influenza virus, adenovirus, parainfluenza virus, and hantavirus) must also be considered. In immunocompromised hosts, the differential diagnosis expands to include a variety of other opportunistic pathogens.

A number of studies have attempted to develop algorithms that can distinguish the causative organism on clinical grounds, but none has provided sufficient predictive value upon which to base pathogen-directed empiric treatment decisions. Hence, a broad differential must be considered upon initial presentation, and definitive diagnosis can be made only with the appropriate microbiologic testing.

  • Complications.Bacteremia, progressive respiratory failure, and hemodynamic instability (ie, septic shock) are the most severe complications of pneumococcal pneumonia. As noted above, patients with either anatomic or functional asplenia are at particularly high risk for sepsis. Spread of infection via either contiguous extension or hematogenous routes constitutes the basis of extrapulmonary complications.

Empyema deserves special attention because it both is common and has severe manifestations if unrecognized. Empyema, which can occur either from hematogenous or contiguous spread, is defined by pleural fluid containing frank pus, a positive Gram stain or positive culture, or a pH of < 7.1. Untreated, it can result in persistent fever and may be the focus for further spread of infection. Rarely, rupture through the chest wall (ie, empyema necessitatis) can occur. If empyema fluid is not properly drained, it may heal with residual fibrosis and result in long-term functional pulmonary defects.

Purulent pericarditis, arthritis, endocarditis, and meningitis are also possible complications.


The diagnostic approach to pneumococcal pneumonia first involves correctly diagnosing the syndrome of pneumonia and second involves defining S pneumoniae as the causative agent. Whether the setting is community acquired or nosocomial, the diagnosis of pneumonia is made by chest radiography of patients with suggestive predisposing factors, symptoms, and physical-examination findings. This avoids the unnecessary costs and medication side effects that are associated with prescribing antibiotics to the large numbers of individuals who have viral upper-respiratory-tract infections and who do not have pneumonia. Although chest radiography has been dismissed by some as being cost ineffective in the outpatient setting, this ignores the incalculable costs of the subsequent development of antibiotic resistance that occurs when antibiotics are inappropriately prescribed. Thus, a chest radiograph is required for all patients in whom pneumonia is suspected.

No constellation of presenting signs or symptoms has proven to be adequately predictive of pneumonia caused by S pneumoniae. Hence, once the diagnosis of pneumonia has been established on clinical and radiographic grounds, definitive diagnosis of pneumococcal pneumonia is based on the identification of the organism in normally sterile fluid (see above and Table 47-1 for microbiologic identification of S pneumoniae). With the rare exception of when transthoracic biopsies are performed, normally sterile fluid means either blood or pleural fluid. As stated above, detection of S pneumoniae in culture of expectorated sputum, although not entirely specific, is sufficient for a probable diagnosis of pneumococcal pneumonia and evidence enough for specific pathogen-directed therapy. The unambiguous sighting of gram-positive diplococci on sputum Gram stain without growth in culture is the least definitive evidence but may be sufficient if other pathogens are excluded and if the patient has an epidemiologic profile consistent with pneumococcal disease.

Obtaining both blood and sputum for culture is the only way to specifically diagnose pneumococcal pneumonia, but there is debate whether to perform these procedures at all, particularly in the outpatient setting. It has been argued that broad-spectrum empiric antibiotics are sufficiently effective in most patients and that specific identification of the pathogen is not required. Until a definitive trial addresses this issue and accounts, by modeling, for the far-reaching effects of broad-spectrum antibiotic use, it is the author's opinion that, when practical to perform, all patients should receive both blood and sputum cultures.

Obtaining cultures can benefit both the individual patient and society. For individuals, determining the specific causative agent allows for altering from empiric broad-spectrum therapy to definitive treatment with an antibiotic that is usually less toxic, expensive, and disruptive to normal flora. In patients who are not responding to initial empiric therapy, a pretreatment culture may be the only opportunity to identify the causative agent and assess its antibiotic resistance patterns. Likewise, the absence of S aureus or gram-negative species in well-collected pretreatment sputum specimens essentially excludes these as pathogens and obviates the need to cover them when reconsidering the antibiotic regimen. For society, in addition to the cost savings derived from cheaper and less toxic antibiotic choices, the benefits of specific pathogen detection include the diminution of antibiotic resistance (and its associated costs) and the epidemiologic surveillance for pathogens such as drug-resistant S pneumoniae and Legionella spp. Knowledge of the prevalence of these pathogens in the community is essential for health care providers as they make empiric treatment decisions for subsequent patients who present with pneumonia.


The approach to treatment of any of the syndromes caused by S pneumoniae must be considered in two parts—empiric and definitive therapy. Empiric therapy is prescribed when a patient presents with a clinical syndrome (eg, pneumonia) and, as is often the case, the causative agent has not yet been identified. Empiric therapy must cover all the epidemiologically likely agents. Definitive treatment is used once S pneumoniae has been identified as the causative agent.

On presentation with a clinical syndrome compatible with pneumonia, most patients are treated empirically with broad-spectrum antibiotics (see Chapter 23). Assuming that a patient is responsive to an empiric regimen that covers S pneumoniae, a switch to definitive therapy depends on the confidence that the clinician and microbiology laboratory have that S pneumoniae is the sole causative agent. As noted above, growth of S pneumoniae in normally sterile body fluids like blood or pleural fluid is definitive proof of pneumococcal disease. In these instances, if other copathogens have been adequately excluded, a switch to definitive therapy is indicated. Growth of S pneumoniae in a culture of expectorated sputum offers suggestive but not definitive proof because some persons may have pneumococcal colonization without pneumococcal disease. The isolated appearance of gram-positive lancet-shaped diplococci on sputum Gram stain (without culture growth) is again suggestive but is also the least definitive. Switching to definitive therapy in these latter two patient groups is a matter of clinical judgment and depends on several factors, including the epidemiologic likelihood of S pneumoniae compared to other pathogens, microbiologic exclusion of other pathogens, patient tolerance of empiric therapy, and cost.

In the past, definitive therapy of pneumococcal pneumonia routinely consisted of penicillin because S pneumoniae was uniformly susceptible. In fact, antimicrobial susceptibility was not routinely tested. This has changed markedly in the past decade. Full susceptibility to penicillin is defined as a minimal inhibitory concentration (MIC) of < 0.06 µg/mL. A recent 30-center survey in the United States found that 14% of isolates had intermediate-level susceptibility (MIC ≥ 0.1–1.0 µg/mL), and 9.5% had high-level resistance (MIC ≥ 2.0 µg/mL). Furthermore, resistance to other antibiotics has also been increasing and is associated with penicillin resistance.

Because of the emergence of resistance, all pneumococcal isolates, including those from expectorated sputum, must undergo antimicrobial susceptibility testing. The choice of definitive therapy must be guided by these results (Box 47-3). Fully susceptible isolates may be treated with parenteral or oral penicillin, depending on the condition of the patient and gastrointestinal function. Fortunately, isolates demonstrating intermediate-level penicillin resistance can still be treated with parenteral penicillin. Because of its excellent bioavailability, amoxicillin is recommended for definitive oral therapy of intermediately resistant strains. Although some highly resistant isolates may also respond to parenteral penicillin or oral amoxicillin, clinical experience in this area is lacking. Hence, agents should be used for which in vitro susceptibility has been shown (eg, ceftriaxone disodium; cefotaxime sodium; a fluoroquinolone such as sparfloxacin or levofloxacin; or vancomycin).

When only a characteristic Gram stain serves as evidence of pneumococcal infection, such isolates should be considered at least intermediately resistant for purposes of definitive therapy unless no such resistance has been reported in the area. Whether these isolates should be considered highly resistant (thus necessitating choices other than parenteral penicillin or oral amoxicillin) again depends on the known prevalence of high-level resistance in the community and whether the empiric regimen to which the patient is responding would be expected to cover isolates with high-level penicillin resistance. If the regimen would not be expected to cover high-level penicillin-resistant isolates and if the patient is nonetheless responding, high-level resistance is unlikely.

Adequate duration of therapy for pneumococcal pneumonia is not known precisely, but patients should be treated for at least 5 days after they become afebrile.


This section focuses on the clinical findings, diagnosis, and treatment of pneumococcal meningitis. Please also see Chapter 52 for a general discussion of meningitis.

Clinical Findings

  • Signs and Symptoms.The clinical presentation of pneumococcal meningitis is similar to other bacterial causes of acute meningitis (Box 47-4). Severe headache, photophobia, neck stiffness, chills, and fever are common. Because of the rapid onset of severe symptoms, most patients present within a short time from the onset of symptoms. Depending on the initiating source of the infection, there may also be pulmonary complaints (see above) or symptoms referable to sinusitis or otitis (see Chapter 9).

Patients with meningitis usually appear extremely ill. Many seek quiet areas with low light because of headache and photophobia. Tachycardia and fever, except in those who have difficulty mounting an elevated temperature (eg, the elderly or those with end-stage renal disease), are almost always present. The hallmarks of meningitis are the presence of one or more of the following three findings: nuchal rigidity, Kernig's sign, or Brudzinski's sign. If patients present within the initial few hours of symptoms, mental status may be normal. If they present later or if treatment has for some reason been withheld, deterioration in mental status and ultimately obtundation may ensue. Papilledema and focal neurologic signs, including cranial neuropathies, may be present especially later in the course of disease. Hyperesthesia (abnormal acuteness of sensitivity to touch) is common. The remainder of the physical examination may reveal evidence of the primary source of infection (eg, purulent discharge from nose or ear, bulging tympanic membrane, or consolidative findings on chest examination). Unlike meningococcal meningitis, a rash is not found.

  • Laboratory Findings.As with pneumococcal pneumonia, an elevated leukocyte count is common in pneumococcal meningitis. Other routine laboratory tests are typically normal. Lumbar puncture, which must be performed as rapidly as the diagnosis of meningitis is entertained, shows increased intracranial pressure (> 180 mm H2O), CSF pleocytosis (from 500 to > 10,000 cells/mm3, with polymorphonuclear-cell predominance), an elevated protein concentration (100–500 mg/dL), and depressed glucose (< 40 mg/dL). Gram stain is often positive for characteristic gram-positive organisms. In contrast to the evaluation of pneumonia, in which a positive Gram stain is not always definitive, the presence of gram-positive diplococci on CSF Gram stain is definitive for the diagnosis of pneumococcal meningitis. In untreated patients, the sensitivity of Gram stain has been estimated to be as high as 90%. CSF culture is positive in most cases, and blood cultures may also show growth. In situations in which antibiotics have been administered before lumbar puncture, both Gram stain and culture may be negative. Cell count, protein, and glucose levels, however, are usually abnormal in these situations and not substantially different from the untreated state. Because these abnormalities are not specific for pneumococcal infection, immunologic methods such as counter immunoelectrophoresis or latex agglutination may be helpful for definitive diagnosis. These techniques provide no additional information to routine Gram stain and culture in the evaluation of CSF samples obtained before therapy.
  • Imaging.Cranial radiography (computed tomography or magnetic resonance scans) is often not indicated in the evaluation of meningitis. In fact, it is contraindicated if it results in undue delays in performing lumbar punctures or initiating empiric therapy. Radiography is indicated when there is clinical evidence of increased intracranial pressure (eg, papilledema) or focal neurologic findings. In these instances, radiography is performed to evaluate whether it is safe to perform a lumbar puncture and to diagnose central nervous system mass lesions, which may require further diagnostic maneuvers. Radiography is rarely useful to confirm positively the presence of meningitis—this is the task of the clinician and the microbiology laboratory. Radiography can be useful in assessing concomitant sinusitis, mastoiditis, or subdural empyema.
  • Differential Diagnosis.The differential diagnosis of acute meningitis is broad and contains both nonbacterial infectious and noninfectious etiologies (see Chapter 52). Once lumbar puncture has been performed and the profile of CSF findings suggests bacterial infection, the differential can be narrowed significantly. The differential diagnosis of acute bacterial meningitis is dependent on age group. In neonates, S agalactiae, members of the family Enterobacteriaceae, and Listeria monocytogenes predominate. With the marked decrease in H influenzae type B infections caused by immunization, S pneumoniae and N meningitidis are now the most prevalent organisms seen in older infants and children. In adults, S pneumoniae, N meningitidis, and L monocytogenes (in older adults) are most common.
  • Complications.The principal immediate and life-threatening complication of meningitis is decreased blood flow to the brain. Rather than a complication, it is best thought of as part of the pathogenic process that occurs if meningitis is unsuccessfully treated. Decreased blood flow may result either from increased intracranial pressure and loss of cerebrovascular autoregulation or from direct vasculitic effects on blood vessels.

BOX 47-3 Definitive Treatment of Streptococcus pneumoniae Pneumonia Based on Penicillin Susceptibility1


Susceptible (MIC <0.1 µg/mL)

Intermediate Resistance (MIC 0.1–1.0 µg/mL)

Highly Resistant (MIC ≥2.0µg/mL)








First Choice

Penicillin V (7 mg/kg orally every 6 h—maximum 500 mg/dose) OR Penicillin G (16, 500 U/kg IV every 4 h—maximum 1 million U/dose)

Penicillin V (500 mg orally four times daily) OR Penicillin G (1 million U IV million every 4 h)

Amoxicillin (13.3 mg/kg orally three times daily— maximum 500 mg/dose) OR Penicillin G (33,000 U/kg IV every 4 h— maximum 2 million U/dose)

Amoxicillin (500 mg orally three times daily) OR Penicillin G (2 million U IV every 4 h)

Cefpodoxime2(5 mg/kg orally every 12 h—maximum 200 mg/dose) OR Ceftriaxone2(25 mg/kg IV every 12 h— maximum 1 g/dose)

Cefpodoxime2(200 mg orally twice daily) ORCeftriaxone2(1 g IV every 12 h)

Second Choice

Amoxicillin (13.3 mg/kg orally three times daily— maximum 500 mg/dose) OR Cefazolin (20 mg/kg IV every 8 h—maximum 1 g/dose)

Amoxicillin (500 mg orally three times daily) OR Cefazolin (1 g IV every 8 h)

Cefuroxime axetil (15 mg/kg orally twice daily— maximum 500 mg/dose) OR Cefazolin (20 mg/kg IV every 8 h—maximum 1 g/dose)

Cefuroxime axetil (500 mg orally twice daily) OR Cefazolin (1 g IV every 8 h)

Cefotaxime2(25 mg/kg IV every 6 h—maximum 1 g/dose)

Cefotaxime2(1 g IV every 6 h)

Penicillin Allergic

Erythromycin (7.5 mg/kg orally every 6 h or 5 mg/kg IV every 6 h—maximum 500 mg/dose) OR Clindamycin 4 mg/kg orally four times daily— maximum 300 mg/dose or 8.3 mg/kg IV every 8 h—maximum 300 mg/dose)

Erythromycin (500 mg orally or IV every 6 h) OR Clindamycin (300 mg orally four times daily or 600 mg IV every 8 h)

Clindamycin (4 mg/kg orally four times daily— maximum 300 mg/ dose or 8.3 mg/kg IV every 8 h—maximum 600 mg/dose)

Clindamycin (300 mg orally four times daily or 600 mg IV every 8 h)

Vancomycin (10 mg/kg IV every 6 h)

Vancomycin (1 g IV every 12 h)

1Choices for oral and parenteral therapy are given. The use of oral vs parenteral therapy depends on the patient's clinical course and gastrointestinal function. All dosages assume normal renal function.
2Assuming in vitro susceptibility. If not susceptible to cephalosporins, a suitable fluoroquinolone (such as sparfloxacin or levofloxacin) or vancomycin must be used.

BOX 47-4 Clinical Findings in Pneumococcal Meningitis




Laboratory Tests1

More Common

· Headache

· Stiff neck

· Fever

· Chills

· Sweats

· Photophobia

· Fever, tachycardia

· Nuchal rigidity

· Kernig's and Brudzinski's signs

· Elevated leukocyte count

· Increased intracranial pressure (>180 mm H2O)

· CSF pleocytosis (500 to 10,000 cells/mm3, with polymorphonuclear cell predominance)

· Elevated CSF protein (100–500 mg/dL)

· Depressed CSF glucose (<40 mg/dL)

Less Common2

· Cough

· Earache or facial pressure/ postnasal discharge

· Confusion

· Hypotension

· Papilledema

· Focal neurologic deficits

· Hyperesthesia

· Altered mental status

· Consolidative findings on chest exam

· Purulent discharge from nose or ear; bulging tympanic membrane

· Hypoxemia

1Radiographic and microbiologic findings are discussed in the text.
2Many of these findings represent the presence of concomitant pneumococcal disease in other organ systems.

BOX 47-5 Definitive Treatment of Streptococcus pneumoniae Meningitis Based on Penicillin Susceptibility1


Susceptible (MIC < 0.1µg/mL)

Intermediate or Highly
Resistant (MIC≥0.1µg/mL)






First Choice

Penicillin G (50,000 U/kg IV every 4 h— maximum 4 million U/dose)

Penicillin G (4 million U IV every 4 h)

Ceftriaxone2 (50 mg/kg IV every 12 h—maximum 2 g/dose)

Ceftriaxone2(2 g IV every 12 h)

Second Choice

Ceftriaxone (50 mg/kg IVevery 12 h— maximum 2 g/dose) OR Cefotaxime (50 mg/kg IV every 6 h; maximum 2 g/dose)

Ceftriaxone (2 g IV every 12 hours OR Cefotaxime (2 g IV every 6 h)

Cefotaxime2 (50 mg/kg IV every 6 h—maximum 2 g/dose)

Cefotaxime2(2 g IV every 6 h)

Penicillin Allergic

Vancomycin3 (15 mg/kg IV every 6 h)

Vancomycin3 (1 g IV every 12 h)

Vancomycin (15 mg/kg IV every 6 h)

Vancomycin (1 g IV every 12 h)

1All dosages assume normal renal function. Adjunctive dexamethasone should be given to children at least 15 min before the first dose of antibiotics. See text for discussion of steroid use in adults.
2Assuming in vitro susceptibility. If not susceptible to ceftriaxone or cefotaxime, vancomycin must be used. Even if isolates are resistant to ceftriaxone and ceftotaxime, one of these agents should be considered for use in addition to vancomycin.
3Either cefotaxime or ceftriaxone may be considered if the penicillin allergy is known not to be anaphylactic immediate type. The penetration of these agents is more reliable than vancomycin.


In patients with compatible clinical findings of meningitis, a lumbar puncture is always indicated except when evidence on imaging suggests a risk of herniation. Although lumbar puncture is critical to diagnose meningitis in general and to detect S pneumoniae specifically, delays in performing it should not delay empiric administration of antibiotics. Blood cultures should also be performed, and it is usually possible to collect them before antibiotic treatment. In such cases in which antibiotics must be given before lumbar puncture, blood cultures may serve as the only definitive means of specific pathogen detection.


As with pneumonia, patients presenting with acute meningitis receive broad-spectrum empiric regimens that should include coverage for S pneumoniae (see Chapter 52). Unlike pneumonia, ascertainment of the causal pathogen is more common in acute bacterial meningitis, thus giving the clinician more frequent opportunities to switch to definitive therapy. Either a characteristic CSF Gram stain or culture for S pneumoniae is definitive proof of pneumococcal meningitis.

Antimicrobial susceptibility testing of S pneumoniae is the most important factor in determining definitive therapy (Box 47-5). Unlike the situation in pneumonia cases, in meningitis cases, isolates that are intermediately resistant to penicillin should not be treated with standard or even higher dosages of penicillin. Instead, susceptibility to ceftriaxone disodium or cefotaxime sodium must be determined. Most isolates that are intermediately resistant to penicillin and some that are highly resistant retain susceptibility to cefotaxime sodium and ceftriaxone disodium. In such cases, these third-generation cephalosporins are the drugs of choice. Vancomycin must be used when there is resistance to both cefotaxime sodium and ceftriaxone disodium. Even if there is resistance to both ceftriaxone and cefotaxime, using one of these agents in addition to vancomycin should be considered. If only a positive Gram stain serves as evidence of pneumococcal meningitis and susceptibility results are not available, it is prudent to consider these isolates as resistant and continue the empiric regimen (if the patient is responding) or change to vancomycin if this was not part of the empiric regimen and the patient is not responding. Antibiotic treatment should be given for 10–14 days.

The use of glucocorticoids, although accepted in children, is still controversial in adults. It may be considered for adults in whom the CSF Gram stain is positive, coma is present, or there is evidence of increased intracranial pressure. Glucocorticoids are best considered in the context of empiric therapy because the first dose should be given before antibiotics.


Primary pneumococcal bacteremia (ie, no identifiable anatomic focus) is common in children but infrequent in adults. S pneumoniae is the major pathogen of otitis media in children, being responsible for between 33% and 50% of all cases in which an etiologic agent can be identified. It is also one of the two most important (along with H influenzae) agents implicated in acute sinusitis. Empyema, typically a complication of pneumonia but rarely a primary diagnosis, has been discussed above. Less common syndromes include endocarditis and pericarditis (see Chapter 11), septic arthritis and osteomyelitis (see Chapter 14), peritonitis (see Chapter 12), endometritis (see Chapter 22), brain abscess (see Chapter 7), and cellulitis (see Chapter 13). Most of these can occur as either a primary infection or as a complication of an initiating clinical focus (eg, pneumonia). (Refer to other chapters in this text for a complete description of these syndromes, including a discussion of empiric and definitive treatments.).

The principles discussed above regarding the various levels of certainty in the attribution of a syndrome to S pneumoniae, knowledge of local prevalence of drug-resistant isolates, and the importance of the antimicrobial susceptibility pattern of the isolate (if obtained) are similarly pertinent when determining empiric and definitive therapy for other syndromes.

Prevention & Control

A polyvalent vaccine against S pneumoniae, manufactured by both Merck (Pneumovax 23) and Lederle (Pnu-Immune 23), contains 25 µg of capsular polysaccaride from each of the 23 most common serotypes responsible for invasive pneumococcal infections in the United States. Vaccine serotypes represent ≥ 85–90% of the isolates responsible for invasive infections. It is important to note that the six serotypes that are the most frequent causes of penicillin-resistant infections are included. The vaccine has been shown to be effective for the prevention of invasive pneumococcal disease (eg, bacteremia with or without pneumonia) in immunocompetent adults aged ≥ 65 years and in persons aged ≥ 2 years with chronic illnesses such as diabetes mellitus, alcoholism, cirrhosis, chronic pulmonary disease, coronary artery disease, and congestive heart failure (Box 47-6). Although convincing data on vaccine effectiveness in other immunocompromised populations (eg, HIV infection in persons from developed countries [data from sub-Saharan Africa show no vaccine efficacy], leukemia, Hodgkin's disease, multiple myeloma, and chronic renal failure) have not been presented, methodologic constraints or frank absence of dedicated studies has precluded excluding a protective role. Therefore, vaccination is generally recommended for these populations.

For infants < 2 years old, for whom the 23-valent vaccine is ineffective, a heptavalent vaccine (Prevnar, Wyeth-Lederle) containing the capsular polysaccharides of serotypes 4, 6B, 9V, 14, 18C, 19F, and 23F conjugated to a nontoxic diphtheria protein has been found effective in the prevention of invasive pneumococcal disease (Box 47-6). A series of immunizations with this conjugate vaccine in infants 2, 4, and 6 months old followed by a fourth dose when these infants are 12–15 months old is now recommended for all newborns. The vaccine is also recommended for high-risk children (eg, those with sickle cell disease or asplenia) between 2 and 5 years of age.

The Advisory Committee on Immunization Practices of the US Public Health Service has recently published updated recommendations on revaccination. This committee does not recommend routine revaccination but suggests that individuals ≥ 2 years old who are at highest risk for serious pneumococcal infection and for whom rapid declines in antibody titers are known to occur should receive a second vaccination 5 years after the initial vaccine. These individuals include patients with HIV disease, functional or anatomic asplenia, leukemia, lymphoma, Hodgkin's disease, multiple myeloma, generalized malignancy, chronic renal failure, nephrotic syndrome, or transplants and those receiving immunosuppressive therapy. It is also recommended that individuals ≥ 65 years old receive a second vaccination 5 years after the initial vaccine, provided that the first vaccine was administered when they were < 65 years old.

BOX 47-6 Control of Invasive Streptococcus pneumoniae Infections


· 23-valent polysaccharide vaccine is recommended for the following persons1:

· Aged ≥65 y

· Aged 2–64 y with chronic illnesses such as chronic cardiovascular disease (coronary artery disease, congestive heart failure, and cardiomyopathies), chronic pulmonary disease (chronic obstructive pulmonary disease but not isolated asthma), diabetes mellitus, alcoholism, chronic liver disease (including cirrhosis), cerebrospinal fluid leaks, and functional or anatomic asplenia (including sickle cell disease)

· Aged 2–64 y living in special environment or social settings, such as Alaskan natives and certain American Indian populations as well as residents of nursing homes and other long-term-care facilities

· Aged 2–64 y who are immunocompromised, including those with HIV infection, leukemia, lymphoma, Hodgkin's disease, multiple myeloma, generalized malignancy, chronic renal failure, or nephrotic syndrome; those receiving immunosuppressive chemotherapy (including corticosteroids); and those who have received an organ or bone marrow transplant
Heptavalent conjugate vaccine is recommended for the following persons:

· Aged <2 y and high-risk children between 2 and 5 years of age
Daily oral penicillin V is recommended for children with functional (eg, sickle cell disease) or anatomic asplenia

Isolation Precautions

None required

1See text for recommendations for revaccination.

In children with either functional (eg, sickle cell disease) or anatomic asplenia, chemoprophylaxis with daily oral penicillin V provides another means of prevention and is also recommended.

Although human-to-human transmission of S pneumoniae is common, illness among contacts is very infrequent in nonepidemic settings. Accordingly, isolation of hospitalized patients or immunization of contacts is not recommended.


Advisory Committee on Immunization Practices: Prevention of pneumococcal disease: recommendations of the Advisory Committee on Immunization Practices (ACIP). Morbid Mortal Wkly Rep 1997;46(RR-8):1. (Evidence-based review of the epidemiology of pneumococcal disease and guidelines for the use of the polyvalent polysaccharide vaccine.)

Advisory Committee on Immunization Practices: Preventing pneumococcal disease among infants and young children: recommendations of the Advisory Committee on Immunization Practices (ACIP). Morbid Mortal Wkly Rep 2000;49(RR-9). (New recommendations on use of conjugate vaccine.)

Afessa B, Greaves WL, Frederick WR: Pneumococcal bacteremia in adults: a 14-year experience in an inner-city university hospital. Clin Infect Dis 1995;21(2):345. (Review of 304 cases of pneumococcal bacteremia from an inner-city hospital in Washington, DC. Case-fatality rates did not change significantly in the past six decades.)

Bartlett JG et al: Community-acquired pneumonia in adults: guidelines for management. Clin Infect Dis 1998;26:811. (Comprehensive guidelines on the diagnosis and management of community-acquired pneumonia from the Infectious Diseases Society of America. Much emphasis is provided on the prominent role of S pneumoniae in this disease.)

Bradley JS, Scheld WM: The challenge of penicillin-resistant Streptococcus pneumoniae meningitis: current antibiotic therapy in the 1990s. Clin Infect Dis 1997;24(Suppl 2):S213. (Thoughtful review of the current predicament faced by clinicians treating pneumococcal meningitis in areas where penicillin resistance is prevalent.)

Breiman RF et al: Pneumococcal bacteremia in Charleston County, South Carolina. A decade later. Arch Intern Med 1990;150(7):1401. (A contemporary reevaluation of the incidence of pneumococcal bacteremia in Charleston County, SC, one decade after the initial sentinel surveillance study was performed. This study is among the most valuable sources for population-based estimates of the incidence and risk factors for pneumococcal bacteremia.)

Doern GV et al: Antimicrobial resistance of Streptococcus pneumoniae recovered from outpatients in the United States during the winter months of 1994 to 1995: results of a 30-center national surveillance study. Antimicrob Agents Chemother 1996;40(5):1208. (A 30-center surveillance study of outpatient isolates from throughout the United States that found that 14.1% of pneumococcal strains exhibited intermediate-level penicillin resistance and 9.5% showed high-level resistance.)

Friedland IR: Comparison of the response to antimicrobial therapy of penicillin-resistant and penicillin-susceptible pneumococcal disease. Pediatr Infect Dis J 1995;14(10): 885. (Prospective observational study of pneumococcal disease (pneumonia, sepsis, and peritonitis—excluding meningitis) in children found that intermediate-level penicillin resistance is of little significance in outcome and that standard β-lactam therapy is still highly effective.)

Henrichsen J: Six newly recognized types of Streptococcus pneumoniae. J Clin Microbiol 1995;33(10):2759. (Description of the six newest S pneumoniae serotypes and an interesting historical overview of the serotyping schemes.)

Jernigan DB, Cetron MS, Breiman RF: Minimizing the impact of drug-resistant Streptococcus pneumoniae (DRSP): a strategy from the DRSP Working Group. JAMA 1996;275(3):206. (Suggestions on how to curb the global emergence of drug-resistant S pneumoniae with emphasis on enhanced surveillance, increased vaccination, and more judicious antibiotic use.)

Tuomanen EI, Austrian R, Masure HR: Pathogenesis of pneumococcal infection. N Engl J Med 1995;332(19): 1280. (Review of the pathogenesis of S pneumoniae-mediated disease with a special emphasis on how the inflammation evoked by the cell wall of S pneumoniae differs from that evoked by endotoxin.)