Hill, James B. MD; Sheffield, Jeanne S. MD; McIntire, Donald D. PhD; Wendel, George D. Jr MD
Acute pyelonephritis is one of the most common medical complications of pregnancy. It occurs in 1–2% of pregnant women and may result in significant maternal morbidity, as well as fetal morbidity and mortality. The clinical course of pyelonephritis in pregnancy was well described 20–30 years ago.1,2 These early studies have defined how we diagnose and manage pyelonephritis today. However, recommendations for screening for asymptomatic bacteriuria during pregnancy,3 newer diagnostic techniques, the development of antibiotic resistance, changing microbial virulence factors, and new antimicrobials may affect diagnosis, clinical course, and management of pyelonephritis today.
We sought to readdress the incidence, risk factors, microbial pathogens, and the clinical complications of pyelonephritis in pregnancy. Said another way, have we been able to alter the frequency and outcomes of pyelonephritis in the past 30 years?
MATERIALS AND METHODS
Beginning January 2000, all pregnant women with acute pyelonephritis admitted to Parkland Hospital, Dallas, Texas, were enrolled in a specialized prenatal clinic for women with infectious diseases. This clinic was staffed by faculty and fellows in the Division of Maternal–Fetal Medicine at the University of Texas Southwestern Medical School. All pregnant women with antepartum pyelonephritis had been treated as inpatients with intravenous antibiotics. All women admitted from January 2000 through December 2001 were included in this study.
The diagnosis of acute pyelonephritis was made on the basis of clinical findings of fever (temperature ≥ 38°C), flank pain, and costovertebral angle tenderness, and of laboratory findings of either bacteriuria (20 bacteria per high-power field) or pyuria.2 Other common symptoms included nausea, vomiting, and chills. The diagnosis was established before the results of urine cultures were known, and all women were evaluated by residents and faculty before admission to the hospital. All of these women were treated with intravenous fluid and antimicrobial agents according to previously described protocols.4 Diagnosis was confirmed by urine culture, from midstream clean catch or urethral catheterization. In most women, a complete blood cell count and serum creatinine were measured. Blood cultures were obtained in the presence of temperature greater than 38.5°C or if the patient had signs of sepsis. Respiratory insufficiency was defined as dyspnea, tachypnea, hypoxemia, and radiological signs of pulmonary infiltrates. Women were admitted to an obstetric extended care unit if they had signs of respiratory insufficiency, clinical signs of sepsis, or multiorgan system dysfunction.5
After recovery from pyelonephritis (afebrile > 24 hours with resolution of symptoms), the women were discharged and seen in the obstetrics infectious disease clinic. All women were placed on urinary suppression with nitrofurantoin, 100 mg daily, for the duration of their pregnancy to prevent recurrent bacteriuria and pyelonephritis.1,2,6 In the majority of women, another urine specimen was obtained for culture after recovery from the acute infection.
Research nurses routinely entered pregnancy outcomes and complications for all women delivered at Parkland Hospital into a computerized database. The data are then assessed for consistency and completeness before electronic storage. Antepartum data on women with acute pyelonephritis were entered into a separate research database and linked electronically to pregnancy outcome data. This analysis excluded cases of recurrent pyelonephritis admission data.
Statistical analysis was performed using SAS 6.12 (SAS Institute, Cary, NC). Comparisons were made with χ2 for categorical data, and Student t test or analysis of variance was used for continuous data. Statistical normality was evaluated using the Shapiro-Wilk statistic. For statistically nonnormal data, Wilcoxon rank-sum and Kruskal-Wallis tests were substituted for Student t test and analysis of variance, respectively. Approval for this study was obtained from the Institutional Review Board of the University of Texas Southwestern Medical Center at Dallas.
During the 2-year study period, we identified 440 cases of acute pyelonephritis in pregnancy; complete delivery data were available for 368 women and their infants. Women delivering at our hospital during the same time period were used for comparison. There were 32,282 deliveries during this period, resulting in an incidence of antepartum acute pyelonephritis of 14 per 1,000 deliveries.
The demographic characteristics of the women with acute pyelonephritis are described in Table 1. The maternal mean age at delivery was less than that of the typical patient in our general obstetric population. There was no difference in ethnic background between women with and those without pyelonephritis. There were significantly more nulliparous women with pyelonephritis (44% versus 37%, P = .003).
As demonstrated in Table 2, the majority (53%) of our cases of acute pyelonephritis occurred in the second trimester. The predominant organism was Escherichia coli, accounting for approximately 70% of cases. Other organisms responsible for infection included Klebsiella-Enterobacter (3%), Proteus (2%), and gram-positive organisms, including group B Streptococcus (10%). Although E coli was the most common uropathogen in the third trimester, there were fewer third-trimester cases caused by E. coli compared with the earlier trimesters (P = .004). Similarly, the frequency of gram-positive uropathogens nearly doubled by the third trimester. We performed a subanalysis of women with positive urine cultures, excluding those women with sterile, unsatisfactory, or unavailable urine results. Overall, 83% of these women had pyelonephritis secondary to E. coli, 11.6% to other gram-positive organisms, predominantly group B Streptococcus, 3.5% to Klebsiella-Enterobacter, and 2.2% to Proteus.
We analyzed the maternal risk factors for pyelonephritis by trimester (Table 3). Overall, 13% of women had a maternal risk factor for antepartum pyelonephritis. The most common risk factor was previous history of pyelonephritis and asymptomatic bacteriuria. For women with type 1 diabetes, the majority of infections occurred in the first trimester (P = .013). When analyzed by trimester, there were no significant differences seen in women with sickle cell disease/trait or prior preterm births. In more than 90% of the cases, women received prenatal care before admission.
The hospital course of women with antepartum pyelonephritis is described in Table 4. One hundred three women (23%, 95% confidence interval [CI] 20–28%) developed anemia, and only 7 (2%, 95% CI 0.6–3%) had transient renal dysfunction. Nearly 1 in 5 women had septicemia when blood cultures were obtained. The average hospital stay was 3.5 days, and the mean length of intravenous antibiotic administration was 3.4 days. Of the 368 women delivered at Parkland Hospital, 19 (5%, 95% CI 3–8%) delivered a preterm infant at less than 37 weeks of gestation. Six of these women delivered at less than 32 weeks. Interestingly, only 4 of the 19 women who delivered at less than 37 weeks delivered during their acute pyelonephritis admission. Twenty-six women (7%, 95% CI 5–10%) delivered an infant weighing less than 2,500 g.
Thirty-two women (7%, 95% CI 5–10%) with acute pyelonephritis developed respiratory insufficiency, and their hospital course is described in Table 5. These women received more intravenous fluids during the first 48 hours of admission than those who did not develop respiratory insufficiency. Their maximum temperature (39.6°C versus 38.5°C, P < .001) and heart rate (129.5 versus 110 beats per minute, P < .001) were significantly higher compared with those who did not develop respiratory insufficiency. More women with respiratory insufficiency developed anemia, with a significantly lower hematocrit (25.3 versus 29.1, P < .001), than those who did not. The incidence of septicemia (28% versus 5%, P < .001) was significantly higher in the respiratory insufficiency group. Escherichia coli was the most frequent urinary isolate in both groups and accounted for more cases of respiratory insufficiency than any other uropathogen. There was no association between respiratory insufficiency and infection with any specific uropathogen.
Twelve women (2.7%, 95% CI 1.4–4.7%) were readmitted for recurrent pyelonephritis. Ten of the 12 women had a positive urine culture. One woman had nephrolithiasis with a negative urine culture. All 12 women were noncompliant with their antimicrobial suppression.
We report a large prospective longitudinal study of a cohort of women hospitalized for acute antepartum pyelonephritis. We used a systematic review of the pregnancies with a codified management scheme of maternal treatment and follow-up.7 We found that the incidence of hospitalization for acute pyelonephritis at our hospital was 1.4%. This incidence is less than the 3–4% rate reported in the 1970s before universal screening for asymptomatic bacteriuria was used.7,8 It is similar to the 1–2% incidence reported with antepartum universal screening.7,8
Gilstrap and colleagues2 at this institution observed that 82% of women with pyelonephritis were less than 25 years of age, 70% were multiparous, and 71% were black women. We also noted an association between acute pyelonephritis and young maternal age in the current cohort. However, we observed no association of pyelonephritis with ethnicity or increasing parity. In fact, we noted the opposite association with parity: pyelonephritis was associated with nulliparity in the contemporary group. In 2000, maternal demographic risk factors for acute pyelonephritis have changed to young age and nulliparity, but not ethnicity.
Acute pyelonephritis is generally described as an infection of late pregnancy and the puerperium. Up to 90% of cases have been reported to occur in the second and third trimesters.2 This is thought to occur because of the increasing urinary tract obstruction with stasis caused by the gravid uterus. In our contemporary cohort, we found that 79% of cases of acute pyelonephritis occurred in the last 2 trimesters of pregnancy, but more than 1 in 5 cases occurred in the first trimester. Overall, we noted more first-trimester cases and fewer third-trimester cases than found in historical reports.
We next examined whether the microbiology of urinary pathogens had changed in our cohort. Among women with positive urine cultures, 83% of the cultures were E. coli. This frequency of infection from E. coli is similar to that reported in 1981. However, the pattern of other microorganisms is different. We noted markedly fewer infections from the Klebsiella-Enterobacter group of organisms (3%) than the 23% noted by Gilstrap and coworkers in 1981.2 Further, we noted a large increase in infection from other organisms, predominantly group B Streptococcus and other gram-positive organisms, accounting for nearly 1 in 8 hospitalized cases of acute pyelonephritis with positive urine cultures.
When the urinary pathogens were examined by trimester of pregnancy, we noted fewer cases of E. coli and more cases caused by other pathogens as pregnancy progresses. In the third trimester, 1 in 4 cases of pyelonephritis was due to an organism other than E. coli. In summary, we noted a preponderance of infection with E. coli, fewer cases from the Klebsiella-Enterobacter group, and more cases caused by gram-positive organisms. This shift in the microbiologic pattern of infection has clinical implications, especially in the third trimester; antimicrobial treatment is chosen empirically based on a presumption of urine culture results.
To investigate the unexpected increased frequency of acute pyelonephritis in the first trimester, we examined the frequencies with common maternal risk factors. We found that only diabetes was more common in the first trimester cases. We were surprised to find no other common risks associated with early acute pyelonephritis in our cohort.
The hospital course of acute pyelonephritis is remarkably similar to that previously reported. However, comparisons may be limited because of changes in available antibiotics and clinical practice patterns. We noted the average hospital stay was 3 days, similar to that reported by Wing et al in 1998.9 Cunningham et al1 also noted that 95% of women hospitalized for acute pyelonephritis were afebrile within 72 hours. Nearly one in 4 women had anemia, similar to the frequency reported in 1991 from this institution.10
Acute renal dysfunction was also uncommon, occurring in only 2% of infected gravidas, markedly lower than the 20% rate noted previously.2,11 We speculate that this may be due to earlier presentation for care and intravenous fluid rehydration. This observation should reassure clinicians using empiric aminoglycoside treatment in women hospitalized with pyelonephritis while awaiting serum creatinine measurements.
High-spiking fevers were common in our cohort and hypoxemia was uncommon. Septicemia complicated nearly 1 in 5 hospitalized cases. In addition, nearly 1 in 10 women required management in an obstetric intensive care unit. Fluid management was aggressive on the first hospital day, with total intake of 3–5 L, or approximately 150 mL/h, compared with nearly 100 mL/h the second day of hospitalization.
Numbers of preterm births and small-for-gestational-age infants were not increased compared with expected rates in our hospital (data not shown). Previous increased rates of adverse pregnancy outcomes were not observed in our cohort, possibly because of improvements in acute care, as well as aggressive follow-up care and antimicrobial urinary suppression.
The association of antepartum pyelonephritis and acute pulmonary injury was first described in 1984.12 Yet, it not uncommonly still complicates this infection.13,14 We noted acute respiratory insufficiency in nearly 1 in 10 women in our cohort. These women were noted to have higher fever, more tachycardia, more tachypnea, more anemia, and more renal dysfunction than those without pulmonary injury. Although they received more intravenous fluid therapy on their first hospital day, it does not appear that they were excessively hydrated. Thus, their acute care does not indicate that acute pulmonary injury was pulmonary edema caused by intravenous fluid overload. Similarly, there was no association with any specific pathogen and acute pulmonary complications.
We believe there are several issues of importance regarding the continued observation that respiratory insufficiency complicates pyelonephritis. Recent reports on the association between genetic polymorphisms and the severity of specific infectious diseases leads us to speculate that genetic variation may predispose certain women to a more complicated course of pyelonephritis. This is being investigated at the molecular level with clinical disease correlation. Furthermore, outpatient management of pyelonephritis in pregnancy should be approached with caution. Although reportedly effective and safe in pregnant women with uncomplicated pyelonephritis,15,16 it must be stressed that 7% of women develop pulmonary insufficiency. One woman in 10 requires admission to an extended care unit, and 17% have septicemia at diagnosis.3
There are several limitations to our study and our conclusions. The entire cohort was admitted for inpatient treatment regardless of signs and symptoms present at diagnosis. Applicability to those centers managing some women with pyelonephritis as outpatients is limited. We are also unable to perform formal statistical comparisons to the historical cohorts referenced. Given these limitations, this serves as a large-cohort clinical description of pyelonephritis in the 21st century in a university setting.
1. Cunningham FG, Morris GB, Mickal A. Acute pyelonephritis of pregnancy: a clinical review. Obstet Gynecol 1973;42:112–4.
2. Gilstrap LC 3rd, Cunningham FG, Whalley PJ. Acute pyelonephritis in pregnancy: an anterospective study. Obstet Gynecol 1981;57:409–13.
3. American College of Obstetricians and Gynecologists. Antimicrobial therapy for obstetric patients. ACOG Technical Bulletin 117. Washington, DC: ACOG; 1988.
4. Cunningham FG, Gant NF, Leveno KJ, Gilstrap LC, Hauth JC, Wenstrom KD. Williams obstetrics. 21st ed. McGraw-Hill: New York (NY); 2001. p. 1256–7.
5. Zeeman GG, Wendel GD Jr, Cunningham FG. A blueprint for obstetric critical care. Am J Obstet Gynecol 2003;188:532–6.
6. Harris RE, Gilstrap LC III. Prevention of recurrent pyelonephritis during pregnancy. Obstet Gynecol 1974;44:637–41.
7. Andrews WW, Cox SM, Gilstrap LC. Urinary tract infections in pregnancy. Int Urogynecol J 1990;1:155–63.
8. Harris RE. The significance of eradication of bacteriuria during pregnancy. Obstet Gynecol 1979;53:71–3.
9. Wing DA, Hendershott CM, Debuque L, Millar LK. A randomized trial of three antibiotic regimens for the treatment of pyelonephritis in pregnancy. Obstet Gynecol 1998;92:249–53.
10. Cox SM, Shelburne P, Mason R, Guss S, Cunningham FG. Mechanisms of hemolysis and anemia associated with acute antepartum pyelonephritis. Am J Obstet Gynecol 1991;164:587–90.
11. Whalley PJ, Cunningham FG, Martin FG. Transient renal dysfunction associated with acute pyelonephritis of pregnancy. Obstet Gynecol 1975;46:174–7.
12. Cunningham FG, Leveno KJ, Hankins GDV, Whalley PJ. Respiratory insufficiency associated with pyelonephritis during pregnancy. Obstet Gynecol 1984;63:121–5.
13. Cunningham FG, Lucas MJ, Hankins GDV. Pulmonary injury complicating antepartum pyelonephritis. Am J Obstet Gynecol 1987;156:797–807.
14. Towers CV, Kaminskas CM, Garite TJ, Nageotte MP, Dorchester W. Pulmonary injury associate with antepartum pyelonephritis: can patients at risk be identified? Am J Obstet Gynecol 1991;164:974–80.
15. Wing DA, Hendershott CM, Debuque L, Millar LK. Outpatient treatment of acute pyelonephritis in pregnancy after 24 weeks. Obstet Gynecol 1999;94:683–8.
16. Millar LK, Wing DA, Paul RH, Grimes DA. Outpatient treatment of pyelonephritis in pregnancy: a randomized controlled trial. Obstet Gynecol 1995;86:560–4.