- Email: [email protected]
An appraisal of treatment guidelines for antepartum community-acquired pneumonia
An appraisal of treatment guidelines for antepartum communityacquired pneumonia Nicole P. Yost, MD,a Steven L. Bloom, MD,a Sherrie D. Richey, MD,b Susan M. Ramin, MD,c and F. Gary Cunningham, MDa Dallas and Houston, Texas, and Anchorage, Alaska OBJECTIVE: The optimal strategy for the initial evaluation and management, including criteria for hospitalization, of pregnant women with pneumonia has not been defined. Our purpose was to evaluate a treatment protocol for antepartum pneumonia and to identify criteria for selection of women for potential outpatient treatment. STUDY DESIGN: A protocol based on British and American Thoracic Society guidelines was introduced and included prompt hospitalization and empiric initiation of erythromycin therapy. Maternal and neonatal outcomes were analyzed to assess the efficacy of the protocol. A second analysis involved the retrospective application of published guidelines to ascertain for which women outpatient management might have been appropriate. RESULTS: There were no maternal deaths among the 133 women studied, and in 14 (10%) women there was a misdiagnosis at admission. Erythromycin monotherapy was judged adequate in all but one of the 99 women so treated. Using a modified version of the American Thoracic Society guidelines, we project that only 25% of the women hospitalized with pneumonia could have been managed safely as outpatients. CONCLUSION: Most pregnant women with pneumonia respond well to monotherapy with erythromycin. Outpatient management may be a reasonable option for selected women. (Am J Obstet Gynecol 2000;183:131-5)
Key words: Pneumonia, pregnancy, erythromycin, outpatient management
Community-acquired pneumonia is one of the more common and serious nonobstetric infections complicating pregnancy. In a review of 209 cases reported since 1989, maternal and perinatal mortality in women with antepartum pneumonia was 3% and 4%, respectively.1 Because of these significant risks to mother and fetus, it has been our practice to hospitalize all pregnant women with pneumonia for close observation and to evaluate their responsiveness to therapy.2 Although hospitalization for all of these women is probably not warranted, there have been no studies that address this issue in pregnancy. Specifically, published guidelines3-5 for the initial evaluation and management of community-acquired pneumonia, including criteria for hospitalization, do not address pregnancy. In 1994 we reported our experiences with 71 pregnancies complicated by pneumonia over the preceding 4 years.6 This report and others7-10 are relatively small, and From the Department of Obstetrics and Gynecology, The University of Texas Southwestern Medical Center, Dallas,a the Providence Alaska Medical Center,b and the University of Texas Medical School at Houston.c Received for publication July 23, 1999; revised December 2, 1999; accepted January 12, 2000. Reprint requests: Nicole P. Yost, MD, Department of Obstetrics and Gynecology, University of Texas Southwestern Medical Center, 5323 Harry Hines Blvd, Dallas, TX 75235-9032. Copyright © 2000 by Mosby, Inc. 0002-9378/2000 $12.00 + 0 6/1/105743 doi:10.1067mob.2000.105743
their disparate therapies do not allow elucidation of optimal treatment strategies. Toward the end of our observational study, we introduced a standardized protocol for the management of pregnant women with communityacquired pneumonia. This protocol was designed by using the guidelines for nonpregnant persons recommended by the British Thoracic Sodiety3 and the American Thoracic Society.4 The primary objective of this study was to evaluate our treatment protocol for pregnant women with community-acquired pneumonia. Although we continued our policy of prompt hospitalization, a secondary aim was to establish criteria that could be applied prospectively to identify pregnant women who could be managed safely as outpatients. Methods Beginning January 1, 1993, as our observational study6 of pneumonia complicating pregnancy was concluding, we introduced a protocol to standardize the management of these women. The guidelines were approved by the institutional review board of the University of Texas Southwestern Medical Center and Parkland Memorial Hospital. As before,6 the diagnosis of pneumonia required typical clinical findings, the absence of another source of infection, and a chest radiograph of a lung infiltrate, as interpreted by a faculty radiologist. On diagnosis, an investigator was notified. Uniform preprinted physician orders were used. All women were 131
132 Yost et al
July 2000 Am J Obstet Gynecol
Table I. Guidelines for hospitalization of nonpregnant patients with community-acquired pneumonia American Thoracic Society4
Fine et al5
Coexisting illness Altered mental status Respirations >30/min Temperature >38.3°C White blood cell count <4000 cells/mm3 or >30,000 cells/mm3 Arterial blood gas with room air—PaO2 <60 mm Hg or PaCO2 >50 mm Hg Serum creatinine >1.2 mg/dL Multilobe involvement or pleural effusion by radiography
Table II. Selected demographic characteristics of women with pneumonia compared with general obstetric population
Factor
Antepartum pneumonia (n = 133)
Race Hispanic 67 (50%) Black 47 (35%) White 17 (13%) Other 2 (2%) Age Mean ± SD (y) 24.5 ± 6.5 ≤15 y 3 (2%) ≥40 y 1 (1%) Nulliparous 43 (32%)
Comparison group Statistical (n = 74,392) significance P < .001 49,699 (67%) 16,614 (22%) 5840 (8%) 2239 (3%) 24.0 ± 5.6 1913 (3%) 579 (1%) 28,490 (38%)
P = .359 P = .818 P = .972 P = .157
admitted to the hospital, and standardized evaluation included cultures of sputum, blood, and urine; hemography; serum creatinine and electrolyte measurement; and arterial blood gas analysis. Treatment included intravenous crystalloid solution for hydration, acetaminophen for fever, and oxygen supplementation when indicated. Empiric antimicrobial treatment was given with intravenous erythromycin (500 mg every 6 hours). Broader antimicrobial coverage was used in women who were known to have sickle cell disease or to be seropositive for the human immunodeficiency virus. The addition of a second-generation cephalosporin in women with bilateral infiltrates was left to the discretion of the admitting physician. Our first analysis of these women was a comparison of their pregnancy outcomes with those of all other pregnant women delivered of their infants at Parkland Hospital within the study period. Selected obstetric and neonatal outcomes for this comparison group were obtained from a previously described computerized database.11 Women with pneumonia were also divided into 2 groups—those with an uneventful hospital course and those with a complicated course, defined as the need for intensive care, spontaneous labor concurrent with symptomatic pneumonitis, readmission, or those with either no improvement or worsening clinical status beyond the first 24 hours of hospitalization.
Coexisting illness Altered mental status Respirations ≥30/min Temperature <35°C or ≥40°C Pulse ≥125 beats/min
Table III. Neonatal outcomes in singleton pregnancies of women with pneumonia compared with general obstetric population
Factor Birth weight Mean ± SD (g) ≤2500 g Gestational age at delivery Mean ± SD (wk) ≤36 wk Stillbirths Neonatal deaths Uncorrected Corrected
Antepartum pneumonia (n = 107)
Comparison group (n = 73,594)
Statistical significance
3129 ± 690 17 (16%)
3278 ± 592 5568 (8%)
P = .028 P < .001
38.3 ± 2.6 14 (13%) 1 (1%)
39.4 ± 2.4 6165 (8%) 430 (0.6%)
P < .001 P = .079 P = .635
0 0
3.9/1000 2.4/1000
P = .517 P = .613
A second analysis was the retrospective application of 2 sets of published guidelines to ascertain whether outpatient management might have been successful. We first applied a modified version of criteria of the American Thoracic Society,4 as shown in Table I. Coexisting illness included chronic obstructive airway disease, including asthma, diabetes, chronic renal failure, heart failure, chronic liver disease, or a postsplenectomy state (sickle cell disease). We also retrospectively applied a modified version of the riskscoring system described by Fine et al5 (Table I). Coexisting medical conditions are specified as neoplastic, cerebrovascular, renal, or liver disease or heart failure. Statistical analyses included the χ2 test for categoric data and the Student t test for continuous variables. P < .05 (2-tailed) was judged to be statistically significant. Data were analyzed by SAS (version 6.12) (SAS Institute, Cary, NC) statistical software. Results In the 5-year period from January 1993 to December 1997, 133 pregnant women were hospitalized with community-acquired pneumonia. A total of 112 (84%) of these women were delivered at our hospital by July 1998. During this same time, 74,392 women were delivered of their infants at Parkland Hospital; thus the incidence of community-acquired pneumonia was approximately 1.5 per 1000 deliveries.
Yost et al 133
Volume 183, Number 1 Am J Obstet Gynecol
Table IV. Clinical characteristics in relation to hospital course Uncomplicated group (n = 96)
Complicated group (n = 23)
Characteristic
No.
%
No.
%
Statistical significance
Respiratory distress History of valvular heart disease Multilobe involvement Diabetes Human immunodeficiency virus positive Asthma Smoker
10 0 13 2 7 26 20
10 0 14 2 7 27 21
12 3 10 3 0 8 6
52 13 43 13 0 35 26
P < .001 P < .001 P = .001 P = .019 P = .182 P = .463 P = .584
The demographic characteristics of the women hospitalized with pneumonia were compared with those of the overall obstetric population (Table II). The percentage of black women with pneumonia was significantly greater than their representation in the overall population (35% vs 22%; P < .001). There were no significant differences in maternal age or parity between the 2 groups. The mean (±SD) gestational age at admission for pneumonia was 24.4 ± 9.2 weeks. Approximately a third of the admissions were in the 2-month winter periods of December and January. It is important that 30% of these 133 women had asthma, and 20% smoked cigarettes. Of the 133 women hospitalized with pneumonia, 48% had a white blood cell count of at least 11,000/µL, and 1% had leukopenia, which was defined as a white blood cell count <4000/µL. Anemia (hematocrit <30%) was present in 25% of the women. Of the 102 women who had arterial blood analyses, 39% had an admission room air PaO2 of <80 mm Hg. Sputum cultures were obtained in 89 women, of which 14 (16%) were positive. The most common organisms isolated were Streptococcus pneumoniae, Haemophilus influenzae, and Staphylococcus aureus. Blood cultures were obtained in 94 women and were positive in only 7%, with S pneumoniae and S aureus being the most commonly isolated pathogens. There were no maternal deaths among these 133 women. Four were admitted to the pulmonary intensive care unit, and 2 of them required mechanical ventilation. One of the latter women was spontaneously delivered of a 33-week stillborn infant while intubated. Neonatal outcomes for women with pneumonia were compared with those of the general obstetric population (Table III). There were 2 abortions and 3 twin gestations in women with pneumonia, but for the purpose of this comparison, only singletons were considered. Infants of women with pneumonia were 150 g smaller in comparison with the overall population (P = .028). Similarly, the frequency of low-birth-weight infants (≤2500 g) was higher in the pneumonia group (16% vs 8%; P < .001). Preterm delivery (≤36 weeks’ gestation) was also more common in women with pneumonia (13% vs 8%); however, this did not reach statistical significance (P = .079).
There was no difference in the perinatal mortality rate between the 2 groups. Fourteen (10%) of the 133 women with pneumonia were misdiagnosed at admission. Their median gestational age was 30 weeks (range, 12-36 weeks). The most common incorrect diagnoses were viral syndrome (n = 4), asthma exacerbation (n = 3), and suspected pyelonephritis (n = 2). Two other women underwent laparotomy for suspected appendicitis. The remaining 3 incorrect diagnoses included sickle cell crisis, gastroenteritis, and endocarditis. The 119 women correctly diagnosed at admission with pneumonia were divided into the following 2 groups: 96 had an uneventful hospital course, and 23 had complications, including requiring intensive care, having spontaneous labor, or requiring readmission and those with no improvement or deterioration in clinical status beyond 24 hours after admission. The clinical characteristics in relation to the hospital course are detailed in Table IV. Women with respiratory distress, those with multilobe involvement, or those with a history of insulin-treated diabetes or valvular heart disease were significantly more likely to have a complicated course. Although women with complicated courses were also more likely to smoke and to have a history of asthma, neither of these risk factors was statistically significant. Finally, as shown in Table V, except for the one stillbirth in the complicated group, neonatal outcomes were similar. Of the 119 women correctly diagnosed with pneumonia at admission, 99 (83%) received erythromycin monotherapy, which was discontinued in 5 women because of gastrointestinal symptoms. Monotherapy was judged inadequate in only 1 woman, who remained febrile after 5 days of erythromycin therapy. This woman had pneumococcal sepsis and later required thoracostomy drainage for an empyema. Of the 20 women not treated with erythromycin monotherapy, 5 were given broader antimicrobial coverage because of human immunodeficiency virus seropositivity or sickle cell disease and 5 were treated with erythromycin and cefuroxime because of bilateral lung infiltrates. The remaining 10 women were inadvertently not treated according to the protocol.
134 Yost et al
July 2000 Am J Obstet Gynecol
Table V. Singleton neonatal outcomes in women with uncomplicated pneumonia compared with those with complications
Factor Birth weight Mean ± SD (g) ≤2500 g Gestational age at delivery Mean ± SD (wk) ≤36 wk Stillbirths Neonatal deaths
Uncomplicated group (n = 85)
Complicated group Statistical (n = 22) significance
3193 ± 560 11 (13%)
2881 ± 989 6 (27%)
P = .169 P = .101
38.4 ± 2.2 9 (11%) 0 (0%) 0 (0%)
38.0 ± 4.0 5 (23%) 1 (5%) 0 (0%)
P = .663 P = .132 P = .048 —
The risk-scoring systems shown in Table I were retrospectively applied to the 119 women correctly identified at admission to have pneumonia. Of the 23 women who had a complicated course, 22 (96%) women would have been hospitalized according to the modified American Thoracic Society4 guidelines; the need for hospitalization of the remaining patient was obvious because of spontaneous labor. By contrast, with the modified criteria suggested by Fine et al,5 only 13 (57%) of the women with a complicated course would have been hospitalized. Of the 119 women admitted with pneumonia, 30 (25%) and 79 (66%) would have been assigned to outpatient care by the American Thoracic Society4 and Fine et al5 guidelines, respectively. Thus, with the American Thoracic Society criteria, the number of women requiring hospitalization would have been reduced by one fourth, and it is important that none of these women had complicated courses. By the guidelines of Fine et al, admissions would have been reduced by two thirds; however, 10 of these likely would have required readmission for complicated courses. Comment Although there were no maternal deaths in this study, pneumonia remains a potentially fatal infection in these otherwise healthy young women.1 Specifically, 2% of these 133 women required tracheal intubation and mechanical ventilation. There was 1 fetal death associated with maternal sepsis, and preterm and low-birth-weight deliveries were almost twice as common in women with pneumonia compared with the general obstetric population. There are 3 important observations that can be made from these pregnant women with community-acquired pneumonia treated according to a uniform protocol. The first is a validation for pregnant women of the treatment guidelines proposed by the American Thoracic Society.4 Specifically, women with uncomplicated disease responded well to hospitalization, respiratory support, and empiric initiation of monotherapy with intravenous
erythromycin. This macrolide antimicrobial was selected because it is considered safe in pregnancy12 and has an excellent spectrum of coverage against common pathogens that cause community-acquired pneumonia (eg, S pneumoniae, Mycoplasma pneumoniae, Chlamydia pneumoniae, and Legionella pneumophila). Indeed, erythromycin remains among the preferred antimicrobial agents for community-acquired pneumonia13 and resulted in cure for all but 1 of 99 women so treated. Overall, the drug was well tolerated, and the newer macrolides have even fewer side effects. A second important finding was that 14 women, or 10% of the overall cohort, had a misdiagnosis at the time of admission. The most striking of these were 2 women who underwent exploratory laparotomy for suspected acute appendicitis. These women were examined at 18 and 30 weeks’ gestation, and it is presumed that anatomic displacement of the appendix contributed to these erroneous diagnoses. Three women were thought to have an uncomplicated asthma exacerbation. We previously reported14 that women with asthma who have fever or clinical findings of a lower respiratory infection should have a chest radiograph to rule out pneumonia. Pyelonephritis was incorrectly diagnosed in 2 women who subsequently had a sterile urine culture. Clearly, pleuritic pain from lower-lobe pneumonia may be confused with costovertebral angle tenderness. A third important observation was that outpatient therapy may be a reasonable option for selected women with antepartum pneumonia. This conclusion arose from retrospective application of published criteria to guide hospitalization versus home therapy. We agree with the American Thoracic Society4 that the decision to hospitalize a patient with community-acquired pneumonia is perhaps the single most important decision during the entire course of illness. When we applied their guidelines for outpatient management, only one fourth of these 119 women would have been assigned to outpatient care; however, all who had a complicated course would have been hospitalized. In contrast, with the modified guidelines of Fine et al,5 two thirds of these women would have been outpatient candidates, but only 57% of those with complications would have been correctly identified for hospitalization. The principal differences between these two guidelines are the inclusion of women with asthma and sickle cell disease and a lower fever threshold in the criteria of the American Thoracic Society (Table I). Our results suggest that these criteria, tempered with sound clinical judgment, may be used to select pregnant women with community-acquired pneumonia who may be appropriate candidates for outpatient management. Yet another practical scheme would be use of the 23-hour observation to determine outpatient treatment versus hospitalization. This has been successfully used in pregnant women with pyelonephritis.15
Yost et al 135
Volume 183, Number 1 Am J Obstet Gynecol
REFERENCES
1. Bloom SL, Ramin S, Cunningham FG. A prediction rule for community-acquired pneumonia [letter]. N Engl J Med 1997;336:1913-4. 2. Cunningham FG, MacDonald PC, Gant NF, Leveno KJ, Gilstrap LC III, Hankins GDV, et al. Williams’ obstetrics. Stamford (CT): Appleton & Lange; 1997. 3. Harrison BDW, Farr BM, Connolly CK, MacFarlane JT, Selkon JB, Bartlett CLR. The hospital management of communityacquired pneumonia. Recommendations of the British Thoracic Society. J R Coll Phys Lond 1987;21:267-9. 4. American Thoracic Society. Guidelines for the initial management of adults with community-acquired pneumonia: diagnosis, assessment of severity, and initial antimicrobial therapy. Am Rev Respir Dis 1993;148:1418-26. 5. Fine MJ, Auble TE, Yealy DM, Hanusa BH, Weissfeld LA, Singer DE, et al. A prediction rule to identify low-risk patients with community-acquired pneumonia. N Engl J Med 1997;336: 243-50. 6. Richey SD, Roberts SW, Ramin KD, Ramin SM, Cunningham FG. Pneumonia complicating pregnancy. Obstet Gynecol 1994;84:525-8. 7. Madinger NE, Greenspoon JS, Ellrodt AG. Pneumonia during
8.
9. 10.
11.
12. 13.
14.
15.
pregnancy: has modern technology improved maternal and fetal outcome? Am J Obstet Gynecol 1989;161:657-62. Berkowitz K, LaSala A. Risk factors associated with the increasing prevalence of pneumonia during pregnancy. Am J Obstet Gynecol 1990;163:981-5. Briggs RG, Mabie WC, Sibai BM. Community-acquired pneumonia in pregnancy [abstract]. Am J Obstet Gynecol 1996;174:389. Munn MB, Groome LJ, Baker SL, Atterbury JL, Hoff C. Pneumonia as a complication of pregnancy [abstract]. Am J Obstet Gynecol 1997;176 (1 Pt 2):S186. McIntire DD, Bloom SL, Casey BM, Leveno KJ. The relationship between birthweight and morbidity and mortality in newborn infants. N Engl J Med 1999;340:123-8. Briggs GG, Freeman RK, Yaffe SJ. Drugs in pregnancy and lactation. Baltimore (MD): Williams & Wilkins; 1998. Bartlett JG, Breiman RF, Mandell LA, File TM Jr. Community-acquired pneumonia in adults: guidelines for management. Clin Infect Dis 1998;26:811-38. Wendel PJ, Ramin SM, Barnett-Hamm C, Rowe TF, Cunningham FG. Asthma treatment in pregnancy: a randomized controlled study. Am J Obstet Gynecol 1996;175:150-4. Millar LK, Wing DA, Paul RH, Grimes DA. Outpatient treatment of pyelonephritis in pregnancy: a randomized controlled trial. Obstet Gynecol 1995;86:560-4.
Bound volumes available to subscribers Bound volumes of the American Journal of Obstetrics and Gynecology are available to subscribers (only) for the 2000 issues from the publisher, at a cost of $122.00 for domestic, $156.22 for Canada, and $146.00 for international for Vol. 182 (January-June) and Vol. 183 (July-December). Shipping charges are included. Each bound volume contains a subject and author index, and all advertising is removed. Copies are shipped within 60 days after publication of the last issue in the volume. The binding is durable buckram with the Journal name, volume number, and year stamped in gold on the spine. Payment must accompany all orders. Contact Mosby, Subscription Customer Service, 6277 Sea Harbor Dr, Orlando, FL 32887. Telephone (800)654-2452 or (407)345-4000. Fax (407)363-9661. Subscriptions must be in force to qualify. Bound volumes are not available in place of a regular Journal subscription.
Key words: Pneumonia, pregnancy, erythromycin, outpatient management
Community-acquired pneumonia is one of the more common and serious nonobstetric infections complicating pregnancy. In a review of 209 cases reported since 1989, maternal and perinatal mortality in women with antepartum pneumonia was 3% and 4%, respectively.1 Because of these significant risks to mother and fetus, it has been our practice to hospitalize all pregnant women with pneumonia for close observation and to evaluate their responsiveness to therapy.2 Although hospitalization for all of these women is probably not warranted, there have been no studies that address this issue in pregnancy. Specifically, published guidelines3-5 for the initial evaluation and management of community-acquired pneumonia, including criteria for hospitalization, do not address pregnancy. In 1994 we reported our experiences with 71 pregnancies complicated by pneumonia over the preceding 4 years.6 This report and others7-10 are relatively small, and From the Department of Obstetrics and Gynecology, The University of Texas Southwestern Medical Center, Dallas,a the Providence Alaska Medical Center,b and the University of Texas Medical School at Houston.c Received for publication July 23, 1999; revised December 2, 1999; accepted January 12, 2000. Reprint requests: Nicole P. Yost, MD, Department of Obstetrics and Gynecology, University of Texas Southwestern Medical Center, 5323 Harry Hines Blvd, Dallas, TX 75235-9032. Copyright © 2000 by Mosby, Inc. 0002-9378/2000 $12.00 + 0 6/1/105743 doi:10.1067mob.2000.105743
their disparate therapies do not allow elucidation of optimal treatment strategies. Toward the end of our observational study, we introduced a standardized protocol for the management of pregnant women with communityacquired pneumonia. This protocol was designed by using the guidelines for nonpregnant persons recommended by the British Thoracic Sodiety3 and the American Thoracic Society.4 The primary objective of this study was to evaluate our treatment protocol for pregnant women with community-acquired pneumonia. Although we continued our policy of prompt hospitalization, a secondary aim was to establish criteria that could be applied prospectively to identify pregnant women who could be managed safely as outpatients. Methods Beginning January 1, 1993, as our observational study6 of pneumonia complicating pregnancy was concluding, we introduced a protocol to standardize the management of these women. The guidelines were approved by the institutional review board of the University of Texas Southwestern Medical Center and Parkland Memorial Hospital. As before,6 the diagnosis of pneumonia required typical clinical findings, the absence of another source of infection, and a chest radiograph of a lung infiltrate, as interpreted by a faculty radiologist. On diagnosis, an investigator was notified. Uniform preprinted physician orders were used. All women were 131
132 Yost et al
July 2000 Am J Obstet Gynecol
Table I. Guidelines for hospitalization of nonpregnant patients with community-acquired pneumonia American Thoracic Society4
Fine et al5
Coexisting illness Altered mental status Respirations >30/min Temperature >38.3°C White blood cell count <4000 cells/mm3 or >30,000 cells/mm3 Arterial blood gas with room air—PaO2 <60 mm Hg or PaCO2 >50 mm Hg Serum creatinine >1.2 mg/dL Multilobe involvement or pleural effusion by radiography
Table II. Selected demographic characteristics of women with pneumonia compared with general obstetric population
Factor
Antepartum pneumonia (n = 133)
Race Hispanic 67 (50%) Black 47 (35%) White 17 (13%) Other 2 (2%) Age Mean ± SD (y) 24.5 ± 6.5 ≤15 y 3 (2%) ≥40 y 1 (1%) Nulliparous 43 (32%)
Comparison group Statistical (n = 74,392) significance P < .001 49,699 (67%) 16,614 (22%) 5840 (8%) 2239 (3%) 24.0 ± 5.6 1913 (3%) 579 (1%) 28,490 (38%)
P = .359 P = .818 P = .972 P = .157
admitted to the hospital, and standardized evaluation included cultures of sputum, blood, and urine; hemography; serum creatinine and electrolyte measurement; and arterial blood gas analysis. Treatment included intravenous crystalloid solution for hydration, acetaminophen for fever, and oxygen supplementation when indicated. Empiric antimicrobial treatment was given with intravenous erythromycin (500 mg every 6 hours). Broader antimicrobial coverage was used in women who were known to have sickle cell disease or to be seropositive for the human immunodeficiency virus. The addition of a second-generation cephalosporin in women with bilateral infiltrates was left to the discretion of the admitting physician. Our first analysis of these women was a comparison of their pregnancy outcomes with those of all other pregnant women delivered of their infants at Parkland Hospital within the study period. Selected obstetric and neonatal outcomes for this comparison group were obtained from a previously described computerized database.11 Women with pneumonia were also divided into 2 groups—those with an uneventful hospital course and those with a complicated course, defined as the need for intensive care, spontaneous labor concurrent with symptomatic pneumonitis, readmission, or those with either no improvement or worsening clinical status beyond the first 24 hours of hospitalization.
Coexisting illness Altered mental status Respirations ≥30/min Temperature <35°C or ≥40°C Pulse ≥125 beats/min
Table III. Neonatal outcomes in singleton pregnancies of women with pneumonia compared with general obstetric population
Factor Birth weight Mean ± SD (g) ≤2500 g Gestational age at delivery Mean ± SD (wk) ≤36 wk Stillbirths Neonatal deaths Uncorrected Corrected
Antepartum pneumonia (n = 107)
Comparison group (n = 73,594)
Statistical significance
3129 ± 690 17 (16%)
3278 ± 592 5568 (8%)
P = .028 P < .001
38.3 ± 2.6 14 (13%) 1 (1%)
39.4 ± 2.4 6165 (8%) 430 (0.6%)
P < .001 P = .079 P = .635
0 0
3.9/1000 2.4/1000
P = .517 P = .613
A second analysis was the retrospective application of 2 sets of published guidelines to ascertain whether outpatient management might have been successful. We first applied a modified version of criteria of the American Thoracic Society,4 as shown in Table I. Coexisting illness included chronic obstructive airway disease, including asthma, diabetes, chronic renal failure, heart failure, chronic liver disease, or a postsplenectomy state (sickle cell disease). We also retrospectively applied a modified version of the riskscoring system described by Fine et al5 (Table I). Coexisting medical conditions are specified as neoplastic, cerebrovascular, renal, or liver disease or heart failure. Statistical analyses included the χ2 test for categoric data and the Student t test for continuous variables. P < .05 (2-tailed) was judged to be statistically significant. Data were analyzed by SAS (version 6.12) (SAS Institute, Cary, NC) statistical software. Results In the 5-year period from January 1993 to December 1997, 133 pregnant women were hospitalized with community-acquired pneumonia. A total of 112 (84%) of these women were delivered at our hospital by July 1998. During this same time, 74,392 women were delivered of their infants at Parkland Hospital; thus the incidence of community-acquired pneumonia was approximately 1.5 per 1000 deliveries.
Yost et al 133
Volume 183, Number 1 Am J Obstet Gynecol
Table IV. Clinical characteristics in relation to hospital course Uncomplicated group (n = 96)
Complicated group (n = 23)
Characteristic
No.
%
No.
%
Statistical significance
Respiratory distress History of valvular heart disease Multilobe involvement Diabetes Human immunodeficiency virus positive Asthma Smoker
10 0 13 2 7 26 20
10 0 14 2 7 27 21
12 3 10 3 0 8 6
52 13 43 13 0 35 26
P < .001 P < .001 P = .001 P = .019 P = .182 P = .463 P = .584
The demographic characteristics of the women hospitalized with pneumonia were compared with those of the overall obstetric population (Table II). The percentage of black women with pneumonia was significantly greater than their representation in the overall population (35% vs 22%; P < .001). There were no significant differences in maternal age or parity between the 2 groups. The mean (±SD) gestational age at admission for pneumonia was 24.4 ± 9.2 weeks. Approximately a third of the admissions were in the 2-month winter periods of December and January. It is important that 30% of these 133 women had asthma, and 20% smoked cigarettes. Of the 133 women hospitalized with pneumonia, 48% had a white blood cell count of at least 11,000/µL, and 1% had leukopenia, which was defined as a white blood cell count <4000/µL. Anemia (hematocrit <30%) was present in 25% of the women. Of the 102 women who had arterial blood analyses, 39% had an admission room air PaO2 of <80 mm Hg. Sputum cultures were obtained in 89 women, of which 14 (16%) were positive. The most common organisms isolated were Streptococcus pneumoniae, Haemophilus influenzae, and Staphylococcus aureus. Blood cultures were obtained in 94 women and were positive in only 7%, with S pneumoniae and S aureus being the most commonly isolated pathogens. There were no maternal deaths among these 133 women. Four were admitted to the pulmonary intensive care unit, and 2 of them required mechanical ventilation. One of the latter women was spontaneously delivered of a 33-week stillborn infant while intubated. Neonatal outcomes for women with pneumonia were compared with those of the general obstetric population (Table III). There were 2 abortions and 3 twin gestations in women with pneumonia, but for the purpose of this comparison, only singletons were considered. Infants of women with pneumonia were 150 g smaller in comparison with the overall population (P = .028). Similarly, the frequency of low-birth-weight infants (≤2500 g) was higher in the pneumonia group (16% vs 8%; P < .001). Preterm delivery (≤36 weeks’ gestation) was also more common in women with pneumonia (13% vs 8%); however, this did not reach statistical significance (P = .079).
There was no difference in the perinatal mortality rate between the 2 groups. Fourteen (10%) of the 133 women with pneumonia were misdiagnosed at admission. Their median gestational age was 30 weeks (range, 12-36 weeks). The most common incorrect diagnoses were viral syndrome (n = 4), asthma exacerbation (n = 3), and suspected pyelonephritis (n = 2). Two other women underwent laparotomy for suspected appendicitis. The remaining 3 incorrect diagnoses included sickle cell crisis, gastroenteritis, and endocarditis. The 119 women correctly diagnosed at admission with pneumonia were divided into the following 2 groups: 96 had an uneventful hospital course, and 23 had complications, including requiring intensive care, having spontaneous labor, or requiring readmission and those with no improvement or deterioration in clinical status beyond 24 hours after admission. The clinical characteristics in relation to the hospital course are detailed in Table IV. Women with respiratory distress, those with multilobe involvement, or those with a history of insulin-treated diabetes or valvular heart disease were significantly more likely to have a complicated course. Although women with complicated courses were also more likely to smoke and to have a history of asthma, neither of these risk factors was statistically significant. Finally, as shown in Table V, except for the one stillbirth in the complicated group, neonatal outcomes were similar. Of the 119 women correctly diagnosed with pneumonia at admission, 99 (83%) received erythromycin monotherapy, which was discontinued in 5 women because of gastrointestinal symptoms. Monotherapy was judged inadequate in only 1 woman, who remained febrile after 5 days of erythromycin therapy. This woman had pneumococcal sepsis and later required thoracostomy drainage for an empyema. Of the 20 women not treated with erythromycin monotherapy, 5 were given broader antimicrobial coverage because of human immunodeficiency virus seropositivity or sickle cell disease and 5 were treated with erythromycin and cefuroxime because of bilateral lung infiltrates. The remaining 10 women were inadvertently not treated according to the protocol.
134 Yost et al
July 2000 Am J Obstet Gynecol
Table V. Singleton neonatal outcomes in women with uncomplicated pneumonia compared with those with complications
Factor Birth weight Mean ± SD (g) ≤2500 g Gestational age at delivery Mean ± SD (wk) ≤36 wk Stillbirths Neonatal deaths
Uncomplicated group (n = 85)
Complicated group Statistical (n = 22) significance
3193 ± 560 11 (13%)
2881 ± 989 6 (27%)
P = .169 P = .101
38.4 ± 2.2 9 (11%) 0 (0%) 0 (0%)
38.0 ± 4.0 5 (23%) 1 (5%) 0 (0%)
P = .663 P = .132 P = .048 —
The risk-scoring systems shown in Table I were retrospectively applied to the 119 women correctly identified at admission to have pneumonia. Of the 23 women who had a complicated course, 22 (96%) women would have been hospitalized according to the modified American Thoracic Society4 guidelines; the need for hospitalization of the remaining patient was obvious because of spontaneous labor. By contrast, with the modified criteria suggested by Fine et al,5 only 13 (57%) of the women with a complicated course would have been hospitalized. Of the 119 women admitted with pneumonia, 30 (25%) and 79 (66%) would have been assigned to outpatient care by the American Thoracic Society4 and Fine et al5 guidelines, respectively. Thus, with the American Thoracic Society criteria, the number of women requiring hospitalization would have been reduced by one fourth, and it is important that none of these women had complicated courses. By the guidelines of Fine et al, admissions would have been reduced by two thirds; however, 10 of these likely would have required readmission for complicated courses. Comment Although there were no maternal deaths in this study, pneumonia remains a potentially fatal infection in these otherwise healthy young women.1 Specifically, 2% of these 133 women required tracheal intubation and mechanical ventilation. There was 1 fetal death associated with maternal sepsis, and preterm and low-birth-weight deliveries were almost twice as common in women with pneumonia compared with the general obstetric population. There are 3 important observations that can be made from these pregnant women with community-acquired pneumonia treated according to a uniform protocol. The first is a validation for pregnant women of the treatment guidelines proposed by the American Thoracic Society.4 Specifically, women with uncomplicated disease responded well to hospitalization, respiratory support, and empiric initiation of monotherapy with intravenous
erythromycin. This macrolide antimicrobial was selected because it is considered safe in pregnancy12 and has an excellent spectrum of coverage against common pathogens that cause community-acquired pneumonia (eg, S pneumoniae, Mycoplasma pneumoniae, Chlamydia pneumoniae, and Legionella pneumophila). Indeed, erythromycin remains among the preferred antimicrobial agents for community-acquired pneumonia13 and resulted in cure for all but 1 of 99 women so treated. Overall, the drug was well tolerated, and the newer macrolides have even fewer side effects. A second important finding was that 14 women, or 10% of the overall cohort, had a misdiagnosis at the time of admission. The most striking of these were 2 women who underwent exploratory laparotomy for suspected acute appendicitis. These women were examined at 18 and 30 weeks’ gestation, and it is presumed that anatomic displacement of the appendix contributed to these erroneous diagnoses. Three women were thought to have an uncomplicated asthma exacerbation. We previously reported14 that women with asthma who have fever or clinical findings of a lower respiratory infection should have a chest radiograph to rule out pneumonia. Pyelonephritis was incorrectly diagnosed in 2 women who subsequently had a sterile urine culture. Clearly, pleuritic pain from lower-lobe pneumonia may be confused with costovertebral angle tenderness. A third important observation was that outpatient therapy may be a reasonable option for selected women with antepartum pneumonia. This conclusion arose from retrospective application of published criteria to guide hospitalization versus home therapy. We agree with the American Thoracic Society4 that the decision to hospitalize a patient with community-acquired pneumonia is perhaps the single most important decision during the entire course of illness. When we applied their guidelines for outpatient management, only one fourth of these 119 women would have been assigned to outpatient care; however, all who had a complicated course would have been hospitalized. In contrast, with the modified guidelines of Fine et al,5 two thirds of these women would have been outpatient candidates, but only 57% of those with complications would have been correctly identified for hospitalization. The principal differences between these two guidelines are the inclusion of women with asthma and sickle cell disease and a lower fever threshold in the criteria of the American Thoracic Society (Table I). Our results suggest that these criteria, tempered with sound clinical judgment, may be used to select pregnant women with community-acquired pneumonia who may be appropriate candidates for outpatient management. Yet another practical scheme would be use of the 23-hour observation to determine outpatient treatment versus hospitalization. This has been successfully used in pregnant women with pyelonephritis.15
Yost et al 135
Volume 183, Number 1 Am J Obstet Gynecol
REFERENCES
1. Bloom SL, Ramin S, Cunningham FG. A prediction rule for community-acquired pneumonia [letter]. N Engl J Med 1997;336:1913-4. 2. Cunningham FG, MacDonald PC, Gant NF, Leveno KJ, Gilstrap LC III, Hankins GDV, et al. Williams’ obstetrics. Stamford (CT): Appleton & Lange; 1997. 3. Harrison BDW, Farr BM, Connolly CK, MacFarlane JT, Selkon JB, Bartlett CLR. The hospital management of communityacquired pneumonia. Recommendations of the British Thoracic Society. J R Coll Phys Lond 1987;21:267-9. 4. American Thoracic Society. Guidelines for the initial management of adults with community-acquired pneumonia: diagnosis, assessment of severity, and initial antimicrobial therapy. Am Rev Respir Dis 1993;148:1418-26. 5. Fine MJ, Auble TE, Yealy DM, Hanusa BH, Weissfeld LA, Singer DE, et al. A prediction rule to identify low-risk patients with community-acquired pneumonia. N Engl J Med 1997;336: 243-50. 6. Richey SD, Roberts SW, Ramin KD, Ramin SM, Cunningham FG. Pneumonia complicating pregnancy. Obstet Gynecol 1994;84:525-8. 7. Madinger NE, Greenspoon JS, Ellrodt AG. Pneumonia during
8.
9. 10.
11.
12. 13.
14.
15.
pregnancy: has modern technology improved maternal and fetal outcome? Am J Obstet Gynecol 1989;161:657-62. Berkowitz K, LaSala A. Risk factors associated with the increasing prevalence of pneumonia during pregnancy. Am J Obstet Gynecol 1990;163:981-5. Briggs RG, Mabie WC, Sibai BM. Community-acquired pneumonia in pregnancy [abstract]. Am J Obstet Gynecol 1996;174:389. Munn MB, Groome LJ, Baker SL, Atterbury JL, Hoff C. Pneumonia as a complication of pregnancy [abstract]. Am J Obstet Gynecol 1997;176 (1 Pt 2):S186. McIntire DD, Bloom SL, Casey BM, Leveno KJ. The relationship between birthweight and morbidity and mortality in newborn infants. N Engl J Med 1999;340:123-8. Briggs GG, Freeman RK, Yaffe SJ. Drugs in pregnancy and lactation. Baltimore (MD): Williams & Wilkins; 1998. Bartlett JG, Breiman RF, Mandell LA, File TM Jr. Community-acquired pneumonia in adults: guidelines for management. Clin Infect Dis 1998;26:811-38. Wendel PJ, Ramin SM, Barnett-Hamm C, Rowe TF, Cunningham FG. Asthma treatment in pregnancy: a randomized controlled study. Am J Obstet Gynecol 1996;175:150-4. Millar LK, Wing DA, Paul RH, Grimes DA. Outpatient treatment of pyelonephritis in pregnancy: a randomized controlled trial. Obstet Gynecol 1995;86:560-4.
Bound volumes available to subscribers Bound volumes of the American Journal of Obstetrics and Gynecology are available to subscribers (only) for the 2000 issues from the publisher, at a cost of $122.00 for domestic, $156.22 for Canada, and $146.00 for international for Vol. 182 (January-June) and Vol. 183 (July-December). Shipping charges are included. Each bound volume contains a subject and author index, and all advertising is removed. Copies are shipped within 60 days after publication of the last issue in the volume. The binding is durable buckram with the Journal name, volume number, and year stamped in gold on the spine. Payment must accompany all orders. Contact Mosby, Subscription Customer Service, 6277 Sea Harbor Dr, Orlando, FL 32887. Telephone (800)654-2452 or (407)345-4000. Fax (407)363-9661. Subscriptions must be in force to qualify. Bound volumes are not available in place of a regular Journal subscription.