Clinical Characteristics and Outcome of Children With Pneumonia Attributable to Penicillin-susceptible and Penicillin-nonsusceptibleStreptococcus pneumoniae
Objective. To compare the clinical characteristics, treatment, and outcome of pediatric patients with pneumonia attributable to isolates of Streptococcus pneumoniae that were either susceptible or nonsusceptible to penicillin.
Design. Multicenter, retrospective study.
Setting. Eight children's hospitals in the United States.
Participants. Two hundred fifty-four children with pneumococcal pneumonia identified from patients enrolled in the United States Pediatric Multicenter Pneumococcal Surveillance Study during the 3-year period from September 1, 1993 to August 31, 1996.
Outcome Measures. Demographic and clinical variables including necessity for and duration of hospitalization, frequency of chest tube placement, antimicrobial therapy, susceptibility of isolates, and clinical outcome.
Results. There were 257 episodes of pneumococcal pneumonia that occurred in 254 patients. Of the 257 isolates, 22 (9%) were intermediate and 14 (6%) were resistant to penicillin; 7 (3%) were intermediate to ceftriaxone and 5 (2%) were resistant to ceftriaxone. There were no differences noted in the clinical presentation of the patients with susceptible versus nonsusceptible isolates. Twenty-nine percent of the patients had a pleural effusion. The 189 (74%) hospitalized patients were more likely to have an underlying illness, multiple lung lobe involvement, and the presence of a pleural effusion than nonhospitalized patients. Fifty-two of 72 hospitalized patients with pleural effusions had a chest tube placed, and 27 subsequently underwent a decortication drainage procedure. Eighty percent of the patients treated as outpatients and 48% of the inpatients received a parenteral second or third generation cephalosporin followed by a course of an oral antimicrobial agent. Two hundred forty-eight of the patients (97.6%) had a good response to therapy. Six patients died; however, only 1 of the deaths was related to the pneumococcal infection.
Conclusion. The clinical presentation and outcome of therapy did not differ significantly between patients with penicillin-susceptible versus those with nonsusceptible isolates ofS pneumoniae. Hospitalized patients were more likely to have underlying illnesses, multiple lobe involvement, and the presence of pleural effusions than patients who did not require hospitalization. In otherwise normal patients with pneumonia attributable to penicillin-resistant pneumococcal isolates, therapy with standard β-lactam agents is effective.
Streptococcus pneumoniae, the most common bacterial pathogen causing community-acquired pneumonia in both adults and children, results in an estimated 500 000 cases of pneumonia each year.1 During the last decade, isolates of S pneumoniae that are resistant to penicillin and other antibiotics have continued to become more frequent.2,,3 The impact that this has had on the treatment of pneumococcal pneumonia is unclear.
The typical clinical presentation and outcome of pneumococcal pneumonia in adults has been well-described; however, there are little data concerning the clinical characteristics of pneumococcal pneumonia in children, especially in the era of increasing antimicrobial resistance.4,,5 Friedland and colleagues,6compared various clinical characteristics in 78 South African children with pneumococcal pneumonia (25 with isolates intermediate to penicillin), and found no difference in the outcome of those children who had penicillin-susceptible versus those with nonsusceptible isolates. Except for small case series or individual case reports, no other studies have focused on the outcome of pneumococcal pneumonia attributable to antibiotic-resistant isolates in children. The purpose of our study was to compare the clinical characteristics, treatment, and outcome of pneumococcal pneumonia in pediatric patients with isolates of S pneumoniae that were either susceptible or nonsusceptible to penicillin.
MATERIALS AND METHODS
The United States Pediatric Multicenter Pneumococcal Surveillance Study Group is comprised of eight children's hospitals in the following cities: Houston, TX; Pittsburgh, PA; Chicago, IL; Columbus, OH; Los Angeles, CA; Little Rock, AR; San Diego, CA; and Winston-Salem, NC. Children with pneumonia attributable to S pneumoniae were identified prospectively from patients with systemic infections enrolled in the United States Pediatric Multicenter Pneumococcal Surveillance Study between September 1, 1993 and August 31, 1996. The diagnosis of pneumococcal pneumonia was based on chest radiograph (CXR) findings and a positive blood and/or pleural fluid culture.
A standardized data form was completed retrospectively for each episode. Information collected included: date of birth; gender; race; date of infection; underlying illness; presenting signs, symptoms, and findings on physical examination; peripheral white blood cell (WBC) count and differential; CXR findings (including the presence of pleural effusion); whether hospitalized; duration of hospitalization; duration of fever and oxygen (O2) requirement for hospitalized patients; chest tube placement and duration; performance of other invasive procedures; antimicrobial therapy; follow-up CXR results; and clinical response. A febrile day was defined as any 24-hour period during which the patient had a temperature ≥100.5°F or 38.1°C. Hypoxia was defined as an oxygen saturation of ≤95% by pulse oximetry (while breathing room air). Tachypnea was defined as a respiratory rate (breaths/minute) greater than the normal range for the respective age groups. For repeat CXRs: resolution was defined as a completely normal CXR with no evidence of abnormality; partial resolution was defined as improvement of the CXR but with persistence of an abnormality; no resolution was defined as no change in or a worsening of the abnormality. Clinical response was defined as good if the patient had improvement or resolution of his/her signs and symptoms during antimicrobial therapy.
Determination of the serotype and minimal inhibitory concentration (MIC) for penicillin and ceftriaxone of each isolate was performed in a central laboratory to which all isolates were sent. Serotyping/serogrouping was performed by quellung reaction, using specific capsular antisera (Statens Seruminstitut, Copenhagen, Denmark; Daco, Inc, Carpinteria, CA). Determination of MIC was done by standard microbroth dilution.7 Guidelines set forth by the National Committee for Clinical Laboratory Standards were used for interpretation of MICs. Susceptibility to penicillin was defined as a MIC ≤0.06 μg/mL, intermediate susceptibility to penicillin as a MIC of 0.1 to 1.0 μg/mL, and resistance to penicillin as a MIC ≥2.0 μg/mL. Susceptibility to ceftriaxone was defined as a MIC ≤0.5 μg/mL, with intermediate susceptibility defined as a MIC = 1.0 μg/mL, and resistance defined as an MIC ≥2.0 μg/mL.8Isolates that are intermediate or resistant are considered nonsusceptible to penicillin or ceftriaxone.
Outcome variables were analyzed by χ2, Fisher's exact test, and Wilcoxon ranked sum test for nonparametric data.
Clinical Characteristics of the Patients
There were 257 episodes of pneumococcal pneumonia that occurred in 254 patients during the 3-year period from September 1, 1993 to August 31, 1996. The patients ranged in age from 0.07 to 311 months (mean, 40.6 months; median, 21.8 months). Fifty-six percent of the patients were male; 45% were African-American, 41% were Caucasian, 10% were Hispanic, and 4% were other races. Ninety-two (36%) of the patients had underlying conditions that included central nervous system disorders (n = 17), genetic problems (n = 12), hematologic conditions (n = 9), prematurity (n = 8), chronic pulmonary disease (n = 8), human immunodeficiency virus (n = 7), cardiac problems (n = 6), immunologic disorders (n = 6), and other miscellaneous problems. Ten (11%) of these patients had received one or more doses of 23-valent pneumococcal vaccine before their infection. Twenty-seven percent of the patients had received a course of antibiotics within the 30 days before developing pneumococcal pneumonia.
The most common presenting signs, symptoms, and findings on physical examination are shown in Fig 1. Eighteen percent of the patients had symptoms compatible with a preceding viral upper respiratory tract infection (median duration, 7 days; range, 1–21 days). The mean duration of fever before diagnosis was 3.1 days, (median, 2 days; range, 0.04–28 days). The median peripheral WBC count at the time of presentation was 19.5 × 103/mm3 (range, 0.1–89.8 × 103 /mm3). Sixty-nine percent of the patients had peripheral WBC counts ≥20 × 103/mm3and 31% had counts ≥30 × 103/mm3. Blood culture was positive for S pneumoniae in 93% of the patients; 7% of the patients had a positive pleural fluid culture alone. Seventy-four percent (n = 189) of the patients were hospitalized for treatment of their pneumonia.
Susceptibility and Serotypes/Serogroups of the Isolates
Of the 257 isolates, 22 (8%) were intermediate to penicillin and 14 (6%) were resistant to penicillin; 7 (3%) were intermediate to ceftriaxone and 5 (2%) were resistant to ceftriaxone. The most commonly recovered serotypes/serogroups, in their relative order of prevalence were 14 (34.4%), 6 (17.2%), 19 (14.4%), 9 (7.2%), 1 (7.2%), 23 (4.8%), 4 (4.8%), and 3 (2.4%).
Inpatient Versus Outpatient Treatment
Table 1 is a comparison of selected clinical and radiographic characteristics of patients who were hospitalized for treatment of their pneumonia and those who were treated as ambulatory patients. Children who were hospitalized were significantly more likely to have an underlying illness (P < .001), multiple lobe involvement (P < .001), and a pleural effusion (P < .001).
For the hospitalized children, the median time to defervescence was 2 days (range, 1–25 days). Seventy-five percent of the patients were afebrile by day 5 of hospitalization and 90% of the patients were afebrile by day 10, regardless of the susceptibility of the pneumococcal isolate to penicillin. The median duration for the supplementation of oxygen in those patients who were hypoxemic was 4 days (range, 0.33–39 days). Seven patients required intubation and mechanical ventilation with the duration ranging from 3 to 21 days. Forty-five (24%) of the patients had an additional imaging procedure (chest ultrasound or computed tomography) performed to delineate further the extent of their pneumonic process.
Seventy-three patients (72 inpatients) had pleural effusions; 48 of these patients initially had thoracentesis performed with 52 of the patients ultimately having a chest tube placed. The median duration of the chest tube placement was 6.5 days (range, 1–21 days). Pleural fluid parameters included a positive Gram stain in 31 (60%) and a positive culture in 35 (67%) of 52 patients. The range of pleural fluid WBC counts was 400 to 275 000/mm3(n = 34) with the percent polymorphonuclear cells ranging from 10 to 100; 28 (82%) of the specimens had ≥50% polymorphonuclear cells. Twenty-two of 28 specimens (79%) had a glucose value of <20 mg/dL and protein values ranged from <2.0 to 5.2 g/dL (n = 29). Pleural fluid lactate dehydrogenase levels ranged from 670 to 63 522 IU/L (n = 18). Urokinase was used as an adjunctive measure in 5 (10%) of the patients with chest tubes; 2 of these patients subsequently had a decortication procedure performed.
Twenty-seven (14.3%) of the hospitalized patients (10.6% of all the patients) underwent a more definitive decortication drainage procedure. The median age of these patients was 61.5 months (range, 4–215 months; mean, 75 months). One-third of these patients had an underlying illness. Ninety-three percent of these patients were febrile before diagnosis of pneumonia with pleural effusion, with a median duration of fever of 6 days (range, 1–20 days; mean, 6.5 days). Decreased breath sounds over the affected area and hypoxia were the most common physical examination findings occurring in 89% of these patients. CXRs showed that 85% had a consolidative pattern with 74% having ≥2 lobes involved. The drainage procedure was performed anytime between the day of admission to 11 days (mean, 4 days; median, 3 days) after hospital admission. The pleural fluid parameters of these patients included pleural fluid WBC counts ranging from 850 to 275 000/mm3(n = 17) with the percent polymorphonuclear cells ranging from 80 to 99. Twelve of 16 specimens (75%) had a glucose value of <20 mg/dL and protein values ranged from 3.1 to 5.1 g/dL (n = 17). There were no significant differences for pleural fluid values (percent glucose concentration <20 mg/dL, percent protein concentration >4.0 g/dL, or percent WBC count >10 000/mm3) between the patients treated with chest tube drainage alone versus those who underwent thoracotomy/decortication. Eighty-five percent of these patients had a positive Gram stain and/or culture or latex agglutination antigen test. After the decortication procedure, the mean time to defervescence was 7 days (range, 1–18 days; median, 6 days), the mean duration of O2 requirement was 5 days (range, 1–39 days; median, 8.8 days), and the mean duration of hospitalization was 13 days (range, 7–176 days; median 23.5 days). Eighty-five percent of these patients had partial resolution of their pneumonic process with the remainder having complete resolution on follow-up CXR performed from 5 to 227 days after the diagnosis of pneumonia.
Penicillin-susceptible Versus Nonsusceptible Isolates
Comparisons of clinical and radiographic characteristics of those patients whose pneumococcal isolates were nonsusceptible to penicillin versus those patients with penicillin-susceptible isolates are shown inTable 2. There were no significant differences for duration of fever, O2 requirement, peripheral WBC count, or CXR findings between the two groups. Furthermore, a comparison of the clinical and radiographic characteristics of only inpatients who had isolates of S pneumoniae that were susceptible to penicillin versus those with isolates of S pneumoniae that were not susceptible to penicillin demonstrated no differences between the two groups.
To determine if the presence of an underlying illness and the susceptibility of the isolate made a difference in the clinical course of the pneumonia in hospitalized patients, selected clinical characteristics of those patients with an underlying illness were compared based on the penicillin susceptibility of their isolate. No differences were demonstrated with regard to age, peripheral WBC count, defervescence, duration of O2 requirement, CXR findings, or need for chest tube. Among patients with an underlying illness, those with pneumococcal isolates that were nonsusceptible to penicillin were hospitalized for a median duration of 12.5 days compared with 7 days in those patients with susceptible isolates, P = .04.
The antimicrobial regimens of the patients were highly variable. However, for those patients treated as outpatients, 80% received a dose of a parenteral second or third generation cephalosporin followed by a course of an oral antimicrobial agent. Seventeen percent of the patients were treated with a course of an oral β-lactam antibiotic alone (penicillins and cephalosporins) and the remaining 3% received a course of a non-β-lactam oral antimicrobial agent.
Forty-eight percent of the inpatients were treated with a course of a second or third generation parenteral cephalosporin followed by a course of an oral antimicrobial agent. Twenty percent of the patients received a parenteral cephalosporin in combination with intravenous penicillin or another intravenous antibiotic, 16% received a parenteral cephalosporin in combination with intravenous vancomycin (median duration, 3 days; mean, 4.7 days; range, 0.33–25 days), and 2% of the patients received intravenous rifampin in combination with other antimicrobial agents. The remaining 14% of inpatients were treated with other intravenous antibiotics alone or in combination with an oral agent or intravenous penicillin, or with monotherapy consisting of a parenteral cephalosporin, or intravenous vancomycin.
Patients With Penicillin Resistant Isolates
Table 3 illustrates selected characteristics of the 16 patients who had pneumococcal isolates that were resistant to penicillin and/or ceftriaxone; only 3 patients received vancomycin. There were no differences found with regard to age, peripheral WBC count, clinical or radiographic findings, antimicrobial regimen, or outcome between these patients and those patients with susceptible isolates.
Follow-up CXRs were performed in 121 (48%) of the 254 patients, ranging from 1 to 274 days after the diagnosis of pneumonia was made. Of these patients, 26% had total resolution and 59% had partial resolution. The susceptibility of the isolate was not related to whether or not resolution occurred.
Response to Therapy
Of the 254 patients, 248 (97.6%) had a good response to therapy and recovered from their pneumonia. The characteristics of the 6 patients (2.4%) who died are shown in Table 4. All of these patients had underlying illnesses, 5 were treated with vancomycin plus a parenteral cephalosporin, however, only 1 of these patients had a penicillin-resistant isolate (MIC = 2.0 μg/mL). Only 1 of the deaths (patient 2) was felt to be related to the pneumococcal infection.
Most of the information pertaining to the clinical presentation and treatment of pneumococcal pneumonia comes from case series and reports in adult patients.4,,59–11 However, only one recently published study in adults has addressed the outcome of pneumococcal pneumonia based on the antibiotic susceptibility of the isolates.12 Pallares et al12 conducted a 10-year, prospective study of 504 adults with culture-proven pneumococcal pneumonia to examine the effect of penicillin and cephalosporin resistance on mortality. Of the 504 patients, 145 (29%) had strains of S pneumoniae that were nonsusceptible to penicillin (62 with MIC ≥2.0 μg/mL) and 31 (6%) had strains that were nonsusceptible to cephalosporins. There were no significant differences in mortality among adult patients with pneumococcal pneumonia based on the susceptibility of their isolate to penicillin or cephalosporins.
Data describing the clinical presentation and treatment of pneumococcal pneumonia in the pediatric population with regard to patients with penicillin nonsusceptible isolates are limited. Friedland and colleagues6 examined selected clinical and radiographic characteristics in 78 South African children who were hospitalized with pneumococcal pneumonia. Twenty-five of these patients had isolates ofS pneumoniae that were intermediate to penicillin; none of the isolates were resistant to penicillin (MIC ≥2.0 μg/mL). The only significant differences found between the two groups were a greater number of children who were human immunodeficiency virus-positive or who had received antibiotic therapy in the month before hospital admission in the group whose isolates of S pneumoniae were nonsusceptible to penicillin.
Hardie et al13 compared the clinical characteristics and outcome of children with complicated parapneumonic effusions caused by S pneumoniae nonsusceptible to penicillin (n = 6) to those cases caused by isolates susceptible to penicillin (n = 17). The results of the study showed that complicated parapneumonic effusions caused by penicillin nonsusceptible S pneumoniae were associated with younger age and higher rates of bacteremia. However, no significant differences in outcome measures were found between patients infected with penicillin-susceptible versus nonsusceptible organisms.
In our group of patients the most common presenting signs and symptoms included fever, a nonproductive cough, and malaise/lethargy. The most common findings on physical examination were decreased breath sounds over the affected area and crackles/rales. When tested, hypoxia (O2 saturation ≤95%) occurred in almost half of the patients. Compared with adult patients (20%–45%), the overall mortality rate for children with bacteremic pneumococcal pneumonia is low. Six patients (2.4%) died but only 1 of the deaths was attributable to the pneumococcal infection.
In adults with pneumococcal pneumonia, pleural effusions may occur in up to 57% of the cases.14,,15 Most pleural effusions will resolve spontaneously without the need for further intervention. However, because of the interplay of a number of host and microbial factors, pleural effusions may progress to become empyemas (a collection of purulent material in the pleural space). Pleural empyema is the most common complication of bacterial pneumonia. Examination of pleural fluid parameters helps to determine the presence of an empyema and the need for a drainage procedure. The occurrence of pleural empyema in pneumococcal pneumonia has an incidence ranging from 2% to 8%.14,,16 Thirty-five of our patients, or 13.8%, had one or more pleural fluid parameters indicative of an empyema.
The most appropriate treatment of pleural empyema in children is controversial. Most children with empyemas can be adequately treated with antibiotics and repeated pleural fluid aspirations or chest tube drainage.17,,18 A more definitive surgical drainage procedure may be required in patients who continue to have respiratory distress, fever, persistent peripheral leukocytosis, and worsening CXR despite the use of appropriate antibiotics and drainage. Although open thoracotomy with lung decortication remains controversial in the treatment of empyemas in children, studies have shown the procedure to be low risk and effective in providing rapid clinical improvement in cases of refractory pleural empyema.19–22 Unfortunately, there have been no controlled prospective trials in children comparing the outcome of different treatment strategies for the treatment of empyema. Thoracoscopy with adhesiolysis and pleural debridement23 and mini-thoracotomy24 are two alternatives to open thoracotomy that also have been used successfully in the treatment of children with refractory empyema. In our series of patients, those who underwent mini-thoracotomy or thoracoscopy (n = 9) for drainage of pleural empyemas did as well as those patients who underwent a more invasive procedure (n = 18).
The optimal antimicrobial regimen for the treatment of infections caused by isolates of S pneumoniae that are nonsusceptible to penicillin, especially those with a high level of resistance (MIC ≥2.0 μg/mL), has not been established. Penicillin therapy is inadequate for the treatment of meningitis attributable to S pneumoniae that is resistant to penicillin.25 However, for penicillin-resistant pneumococcal infections outside of the central nervous system, the data are not clear. Although most experts recognize that penicillin therapy may be effective for infections caused by strains with intermediate susceptibility to penicillin,26–28 there are little or no data pertaining to isolates resistant to penicillin. The achievable peak serum concentrations for standard doses of oral β-lactam agents (eg, amoxicillin with or without clavulanic acid or cephalosporins, etc) may be 3 to 7 times greater than the MIC of a resistant strain.29,,30 Although some studies suggest that infections attributable to isolates of S pneumoniae that are intermediate as well as some that are resistant to penicillin may respond adequately to penicillin and other β-lactam agents,6,,10,28,31 therapy with the extended-spectrum cephalosporins and non-β-lactam agents (vancomycin) is usually recommended for patients with isolates of pneumococci that have a high level of resistance to penicillin or those who are critically ill. Therapeutic changes have usually been based on the susceptibility of the isolate and not on the clinical response of the patient.
Among our patients with resistant isolates, 1 patient whose isolate had an MIC to penicillin of 4 μg/mL was treated as an outpatient with a 10-day course of amoxicillin/clavulanic acid. Another patient whose isolate had an MIC to penicillin of 4 μg/mL and an MIC to ceftriaxone of 8 μg/mL was treated as an outpatient with a single intramuscular dose of cefuroxime of 40 mg/kg followed by a 14-day course of amoxicillin. Both patients recovered from their pneumonias uneventfully despite having isolates with susceptibilities that were in the highly-resistant range. All of the other patients with highly-resistant isolates received at least one dose of an extended spectrum β-lactam agent. Two patients also received vancomycin in combination with an advanced generation cephalosporin and 1 patient each was treated with rifampin and vancomycin and rifampin in combination with an advanced generation β-lactam agent. At recommended doses, the peak serum concentrations of these agents can range from 100 to 200 μg/mL, therefore providing blood and tissue concentrations that may exceed the MIC of a resistant isolate by 100-fold.32–34 Scaglione and colleagues35 examined the distribution pharmacokinetics of cefotaxime and ceftriaxone into serum and pleural exudates of adults who were given a single intravenous 1-gram dose. Very high levels of the free drug are achieved both in serum and in pleural exudate.
The best treatment regimen for pneumococcal pneumonia is still unknown. Clinical outcome after treatment of pneumococcal pneumonia in children needs further investigation. Our data suggests that in otherwise normal patients with pneumonia attributable to penicillin-resistant pneumococcal isolates, therapy with standard β-lactam agents is effective; however, the issue of optimal therapy will have to be continually reassessed as the frequency and intensity of S pneumoniae resistance to various antimicrobial agents continues to increase.
This study was supported in part by a grant from Roche Laboratories.
We thank Constance Rothermel, PhD, for her support. We also acknowledge the help of the following individuals: Timothy Postula; Tracye Paris, RN; Susana Aragon, RN; Michelene Ortenzo; Nancy C. Tucker, RN; and Kathyann Marsh, MSN.
- Received December 17, 1997.
- Accepted May 13, 1998.
Reprint requests to (T.Q.T.) Assistant Professor of Pediatrics, Northwestern University Medical School, 2300 Children's Plaza, Box 20, Chicago, IL 60614.
- CXR =
- chest radiograph •
- WBC =
- white blood cell •
- MIC =
- minimal inhibitory concentration
- Appelbaum PC
- ↵National Committee for Clinical Laboratory Standards. Performance Standards for Antimicrobial Susceptibility Testing. Sixth informational supplement M1150–56. Wayne, PA: National Committee for Clinical Laboratory Standards; 1995
- ↵Minimum Inhibitory Concentration (MIC) Interpretive Standards (μg/mL) for Streptococcus spp. Table 2 C. M100-S7. NCCLS Vol 17. No 2. 1997
- Hardie WD,
- Roberts NE,
- Reising SF,
- et al.
- Bryant RE,
- Salmon CJ
- McLaughlin FJ,
- Goldmann DA,
- Rosenbaum DM,
- et al.
- American Academy of Pediatrics, Committee on Infectious Diseases
- Ginsburg CM,
- McCracken GH Jr.,
- Nelson JD
- Krause PJ,
- Owens NJ,
- Nightingale CH
- Tan TQ,
- Mason EO,
- Kaplan SL
- Trang JM,
- Jacobs RF,
- Kearns GL,
- et al.
- Scaglione F,
- Raichi M,
- Fraschini F
- Copyright © 1998 American Academy of Pediatrics