Managing Nonresolving Pneumonia

Natural History of Pneumonia

Pneumonia that fails to respond to treatment is a common problem. Although quantification of the frequency of this problem is difficult, approximately 15% of pulmonary consultations and 8% of bronchoscopies are done to evaluate nonresolving pneumonia. In the intensive care unit (ICU), up to 90% of patients will have persistent radiographic infiltrates on chest x-ray (CXR). Clinicians are confronted with a complex challenge when this occurs.

It is difficult to define and recognize:

• Normal resolution
• Delayed resolution
• Progression of disease.

A specific pathogen cannot be identified in up to 50% of cases of community-acquired pneumonia (CAP), and two or more etiologies are identified in 5% to 40% of cases. This leads to significant diagnostic uncertainty when patients fail to respond to empiric therapy. As a result, when pneumonia fails to respond to treatment, the question becomes whether the diagnosis of pneumonia is even correct, since many conditions can mimic pneumonia. Moreover, pneumonia that is slow to resolve or fails to respond is influenced by many factors in several domains, including virulence of pathogenic agent, host defenses of patient and, more recently, antimicrobial resistance.

The first step in evaluating nonresolving pneumonia is thus to discriminate between normal and nonresolving pneumonia in order to avoid unnecessary diagnostic tests. This requires an understanding of the pathophysiology of pneumonia resolution. Once nonresolving pneumonia is identified, the second step is a systematic consideration of infectious and noninfectious etiologies. Finally, an approach to diagnosis and management must be developed based on the differential diagnosis.

Pathophysiology

The difficulty in defining the nonresolving pneumonia syndrome is that the normal resolution of pneumonia is not a clearly defined process. Given this uncertainty, it is useful to consider the resolution of pneumonia as a spectrum including:

• Normal resolution
• Slowly resolving pneumonia
• Progressive pneumonia.

The parameters used to describe this spectrum include both clinical and radiographic criteria. Clinical criteria that have been studied include (Table 9-1):

• Fever
• Cough
• Crackles
• White blood count (WBC) count
• Partial pressure of arterial oxygen (PaO₂) level
• C-reactive protein (CRP).

Subjective response is usually noted within 3 to 5 days of starting treatment. Fever resolves within 3 days and leukocytes within 4 days. Fatigue is present in at least 50% of CAP patients at 4 weeks. Most studies of resolution of pneumonia have not focused on symptoms but on radiographic resolution. Normal resolution is differentiated from slowly resolving pneumonia. Investigators differ slightly in their definition of slow resolution but in most instances, slowly resolving pneumonia has been defined by the persistence of radiographic abnormalities for >1 month in a clinically improved host.

More recently, investigators have begun to examine symptom resolution in convalescence of patients who have been hospitalized and have noted that cough and fatigue may persist for many weeks, if not months, after the index episode of illness. Although the microbiologic causative agent may have been eradicated and radiologic resolution has occurred, patients may not return to baseline health status, particularly if concomitant illness is present. Some patients suffer decrements in quality of life and impairment in functional ability. Hence, in certain patients, notably older adults, the natural history of even “normally” resolving pneumonia may be viewed as a prolonged and debilitating illness with significant extrapulmonary sequelae far beyond infection in the distal lung.

From a clinical standpoint, the critical distinction lies in differentiating pneumonias that are progressive from those that are merely slow to resolve. The latter can usually be observed without further testing, while the former require further investigation. Regression of pneumonia must take into account factors that affect the expected rate of resolution. These factors include:

• Comorbidities
• Age
• Severity
• Type of infectious agent.

Enlarge 

Comorbidities and Resolution of CAP

Pneumonia frequently occurs in patients with comorbidities that may delay its resolution (Table 9-2). While patients without these comorbidities will usually demonstrate clearing of radio­graphic infiltrates by 4 weeks, only 20% to 30% of patients with these comorbidities will clear by 4 weeks. The frequency of these comorbidities increases with age and thus concurrent comorbidities are more common in the elderly. As an example, in patients <50 years of age, COPD is present in 5% of CAP cases. In patients >50 years of age, COPD is present in >30% of CAP cases.

Enlarge 

Age and Resolution of CAP

Despite the concurrence of comorbidities and advanced age, several studies have demonstrated that age itself is an independent risk factor for delayed clearing. Approximately 90% of patients <50 years of age show radiographic resolution by 4 weeks. In contrast, only 30% of patients >50 years of age without concurrent disease will have radiographic resolution by 4 weeks. Thus age is among the most important factors associated with delayed resolution. With increasing life expectancy, it may be expected that the scope and impact of aging on resolution will only increase.

Severity and Resolution of CAP

In addition to being associated with comorbidities, age is also associated with an increased risk for more severe pneumonia. However, severity of disease remains an independent risk factor for delayed resolution and time to stability. The time to radiographic resolution for severe CAP has been estimated at 10 weeks, as compared with 3 to 4 weeks for mild-to-moderate pneumonia. Definitions of normal resolution vary significantly in the literature, in part because of wide variations in the severity of illness triggering admission. This has an impact on the expected normal rate of resolution, since severity impacts on time to initial stabilization and the rate of resolution.

Infectious Agents and Resolution of CAP

The rate of radiographic and clinical improvement also varies with the infectious agent (Table 9.3). This section will focus on the features that are relevant to resolution of pneumonia with respect to the most common microorganisms. These include S pneumoniae, C pneumoniae, H influenzae and Legionella and Mycoplasma spp.

Enlarge 

S pneumoniae

Pneumococcal pneumonia is still a significant cause of CAP infections, and therefore accounts for many infectious nonresolving pneumonia syndromes. It is also the best studied of the infectious etiologies in terms of the rate of resolution and the factors that affect resolution. In normal individuals without predisposing illnesses, clinical improvement precedes radiographic improvement.

Clinical improvement is relatively rapid in uncomplicated cases. When auscultation was the primary tool for assessing response to therapy, clinicians were able to detect abnormal findings on physical examination in only 8% of patients at 1 month. Similarly, fever resolves rapidly, with only 6% of patients demonstrating fever beyond 20 days. Risk factors for delayed resolution of auscultatory findings and fever include more severe presentation and multilobar disease.

In contrast, radiographic improvement is often much slower. Despite relatively rapid clinical improvement, anywhere from 20% to 30% of patients will have no improvement on CXR at 1 week. Indeed, initial worsening of the CXR is common. Risk factors for delayed radiographic resolution include:

• Bacteremia
• Persistent fever or leukocytosis beyond 6 days
• Age >50 years
• COPD
• Alcoholism.

Radiographic clearing occurs in 1 to 3 months in nonbacteremic cases and in 3 to 5 months in bacteremic cases. Residual radiographic abnormalities are rare in nonbacteremic cases but are present in up to 35% of bacteremic cases.

L pneumophila

Legionella is increasingly recognized as an important pathogen in patients with severe CAP. Legionella is one of the three most frequent etiologic agents that cause rapidly progressive pneumonias. It is frequently encountered in the compromised host and in the elderly. Established risk factors for infection with Legionella include:

• Cigarette smoking
• Alcoholism
• Age >65
• Immunosuppression, corticosteroids, diabetes
• Renal disease
• Bone marrow transplantation
• HIV.

Many of these predisposing conditions are likewise risk factors for delayed resolution, so it is not surprising that the rate of resolution for Legionella infection is slower than that for other organisms.

Ninety percent of Legionella infections are due to L pneumophila, and 80% of these are due to serogroup 1. Most of the literature on the natural history of Legionella infections is thus based on this one serogroup. As in pneumococcal infections, clinical improvement precedes radiographic improvement. The radiographic infiltrates and clinical picture are usually indistinguishable from severe pneumococcal infections. There is usually an initial patchy infiltrate that subsequently becomes confluent and even bilateral, often despite appropriate antibiotic therapy.

The distinguishing features of Legionella infections are the:

• Propensity for initial radiographic deterioration
• Prolonged resolution of radiographic infiltrates
• Prolonged convalescence associated with this infection.

Radiographic deterioration occurs in up to two thirds of patients infected with Legionella, as compared with 4% of patients with nonbacteremic pneumococcal pneumonia. In addition, after this initial deterioration, resolution is slower than with pneumococcal infections. Radiographic clearing only begins after 2 to 3 weeks, with 50% being abnormal at 10 weeks. Resolution may take as long as 6 to 12 months, with residual fibrosis being evident in up to 25% of patients. Even after radiographic resolution, generalized weakness and fatigue may persist for months. In the initial description of Legionnaires’ disease in Philadelphia, patients frequently complained of fatigue and shortness of breath when surveyed up to 2 years after the event, with half demonstrating residual abnormalities on pulmonary function testing.

M pneumoniae

M pneumoniae is a common cause of respiratory tract infections; however, it is a relatively rare cause of severe pneumonia. Clinically apparent pneumonia occurs in only 3-13% of patients infected, with most patients being young adults. Mycoplasma accounts for approximately 5% of hospital pneumonias but is unusual in those >65 years of age. Because these infections generally are less severe and occur in a younger population, it is not surprising that the rate of resolution is faster than with other types of pneumonia.

The initial radiographic pattern is one of interstitial infiltrates, with progression to air space disease with consolidation. Multilobar involvement is common, occurring in 50% to 60% of cases. Radiographic deterioration on treatment is rare, occurring in <25% of cases. ARDS is a rare complication of Mycoplasma pneumonia.

In contrast to Legionnaires’ disease, rapid resolution of Mycoplasma pneumonia is common. There is usually a rapid clinical improvement that occurs in the first 2 weeks, in part reflecting the predominantly young population affected. Average duration of radiographic abnormalities is 2 to 4 weeks, depending on the use of antibiotics. Forty percent have complete radiographic resolution at 4 weeks and 90% at 8 weeks. Residual scarring and fibrosis are rare.

C pneumoniae

C pneumoniae infection is common, with 30-50% of young adults having serologic evidence of infection. Distinguishing features of Chlamydia infection include:

• Increased frequency of hoarseness
• Lack of fever
• Prolonged period before seeking medical attention
• Extrapulmonary manifestations, including:
  • Erythema nodosum
  • Encephalitis
  • Guillain-Barré syndrome.

The disease is relatively mild and mortality is rare, with prompt resolution as the rule in younger patients. However, relapse is common. Radiographically, Chlamydia pneumonia is indistinguishable from other pneumonias, with lobar and interstitial infiltrates being common. Initial radiographic deterioration is rare, with radiologic clearing requiring 1 to 3 months. Resolution is intermediate between Mycoplasma and Legionella. Fifty percent of chest radiographs clear by 4 weeks, with up to 20% taking >9 weeks. Residual radiographic scarring and fibrosis are seen in 10% to 20% of cases.

H influenzae

H influenzae has become an increasingly common cause of pneumonia and is now recognized as a common pathogen in the elderly, in hospitalized patients, and in cigarette smokers. Risk factors for severe infection include:

• COPD
• Malignant disease
• Diabetes
• Alcoholism
• Immunosuppression.

In a surveillance study in Finland (Takala 1990), 71% of cases occurred in patients who were severely immunocompromised and 55% of invasive cases occurred in those over the age of 50. Invasive cases are more commonly caused by encapsulated strains, which are also associated with a higher risk for severe sepsis and mortality. Unencapsulated strains are less frequently associated with mortality but are more often associated with a prolonged febrile tracheobronchitis.

The clinical presentation of pneumonia caused by Haemophilus is not unique and it is therefore impossible to reliably differentiate it from other pneumonias, particularly that caused by pneumococcus. A multi­lobar pattern of bronchopneumonia with a pleural effusion is considered classic, but this finding is by no means specific.

The natural history of Haemophilus infection has not been well studied, and there are no distinguishing features regarding the rate of resolution. Based on its propensity to infect the immunocompromised and elderly, the rate of resolution can be expected to be slow. Clinical improvement is also slow, with many patients being hospitalized for 2 to 3 weeks, with only half returning to their previous level of function by 6 weeks. Similarly, radiographic resolution can be expected to be slow relative to that of other pneumonias.

Pneumonia of Unknown Etiology

Because S pneumoniae and L pneumophila are both common in severe CAP, the normal resolution time for severe CAP may be expected to range from 3 to 12 weeks. Since half of all pneumonias will have no isolated pathogen, it becomes clear that the possible upper limit of normal resolution will be quite high.

Based on these studies, it is apparent that the normal time to resolution for severe CAP has a broad distribution curve, depending upon a variety of factors. Many patients with nonresolving pneumonia will actually be within the limits of normal resolution once these other factors are taken into consideration. Patients with slow radiographic resolution but a good clinical response can be defined as having slowly resolving pneumonia.

At some point in this spectrum, however, the patient crosses into the area of nonresolving pneumonia. Those patients with clinical deterioration under therapy can be defined as having progressive pneumonia. These two categories have significant overlap but are useful clinical definitions since patients with progressive disease are more likely to warrant additional diagnostic testing. Importantly, the term “pneumonia” in this situation does not necessarily equate with infection, since many patients with clinical deterioration may have a noninfectious disorder. Progressive disease in these cases may be due to factors associated with either infectious or noninfectious etiologies.