Keywords
asymptomatic bacteriuria - acute cystitis - nitrofurantoin - pyelonephritis - trimethoprim-sulfamethoxazole
- uropathogen - urinary tract infections
Lower urinary tract infections (LUTIs) are one of the most common medical complications
of pregnancy, occurring in 8% of U.S. women.[1] LUTIs include both asymptomatic bacteriuria (ASB) or acute cystitis (AC). Left untreated,
ASB develops into symptomatic AC in 30% of patients and may progress to pyelonephritis
in up to 50% of those patients.[2] It is estimated that 0.5 to 2% of all pregnant women are hospitalized for treatment
of pyelonephritis.[3]
[4] Pyelonephritis is associated with increased risk of anemia, sepsis, acute pulmonary
insufficiency, acute renal dysfunction, and spontaneous preterm birth (PTB).[3] Given the increased maternal morbidity of untreated LUTI, the United States Preventive
Services Task Force recommends routine screening for and treatment of ASB and AC during
pregnancy.[5]
Treatment of ASB and AC in pregnancy must cover the most common uropathogens, considering
the altered pharmacokinetics, and must take into account teratogenicity and fetal
toxicity of some antibiotics during pregnancy. The most common urinary tract pathogen
is Escherichia coli, accounting for 80 to 90% of UTIs in pregnancy.[1] Other common UTI pathogens include Group-B Streptococcus, Staphylococcus saprophyticus, Klebsiella pneumoniae, and Proteus mirabilis. Many antibiotics offer sufficient coverage of these microorganisms; however, antibiotic
toxicity in pregnancy restricts treatment options. Antimicrobial agents currently
approved for use in pregnancy include nitrofurantoin, β lactams, cephalosporins, and
fosfomycin trometol.[6] However, current American College of Obstetricians and Gynecologists recommendations
identify nitrofurantoin and sulfonamides, primarily trimethoprim-sulfamethoxazole
(SMZ-TMP), as first-line therapies for LUTI in pregnancy.[7]
Despite these recommendations, a recent CDC study that used MarketScan Commercial
Database to find pregnancies with UTIs and filled antibiotic prescriptions, and included
450,000 pregnancies of which nearly 40,000 met inclusion criteria, found that less
than half of LUTI were treated with nitrofurantoin or SMZ-TMP.[8] These data suggest nonfirst-line antibiotics are commonly used to treat LUTI in
pregnancy. However, the efficacy in preventing progression to pyelonephritis of nonfirst-line
antibiotics for UTI in pregnancy is unknown. The purpose of this study is to compare
the rates of progression to pyelonephritis and obstetric outcomes among women treated
with first-line antibiotic therapies, specifically nitrofurantoin or SMZ-TMP, with
those treated with nonfirst-line therapies.
Materials and Methods
This was a retrospective cohort study of all pregnant women presenting at a single
health system with LUTI, including ASB or AC, during pregnancy. Women delivering at
a Duke University-affiliated hospital from July 1, 2013 to May 1, 2019 were included.
Eligible patients were identified using International Classification of Diseases,
9th (ICD-9) and 10th (ICD-10) Revision codes for ASB, acute cystitis, and pyelonephritis.
Women were included if they were diagnosed with LUTI during pregnancy and delivered
during the study period in the health care system. Women with missing culture data,
antibiotic data, or delivery information were excluded. Women with documentation of
medication nonadherence per provider records were also excluded. ASB and AC were defined
as urine culture positive for a single organism ≥105 cfu/mL without symptoms and ≥102 cfu/mL with symptoms, respectively.[9] Pyelonephritis was defined by clinical diagnosis documented by the provider during
admission.
The electronic medical record was reviewed, and demographic variables, antepartum
and pregnancy complications, delivery, and postpartum data were abstracted by trained
chart abstractors. The exposure of interest was receipt of a “first-line” antibiotic
for the treatment of LUTI. First-line antibiotic therapy was defined as nitrofurantoin
or SMZ-TMP.[10] Nonfirst-line therapy included all other antibiotics. Antibiotic choice was at the
discretion of the provider.
The primary outcome was progression to pyelonephritis. Secondary outcomes included
length of antibiotic course, pyelonephritis-related anemia (defined as the hematocrit
nadir during the admission for pyelonephritis), sepsis, pyelonephritis length of stay,
preterm delivery, and low birth weight. Preterm delivery was defined as delivery prior
to 37 weeks. Short course of antibiotics was defined as less than 7 days. Maternal
anemia was defined as a hematocrit nadir less than 32%.[11] Women who received first-line antibiotic therapy (FLT) were compared with those
who received nonfirst-line therapies (nFLTs).
Baseline demographics were analyzed with Student's t-test, Wilcoxon rank sum tests, Chi-square tests, or Fisher's exact tests as appropriate.
Multivariate logistic regression was used to determine significant predictors of the
primary outcome. Statistical significance was defined as p < 0.05. Statistical analysis was performed using STATA software version 14.0. This
study was approved by the Duke University IRB.
Results
Of 722 women with LUTI, 476 women met inclusion criteria with culture-confirmed LUTI
and documentation of antibiotic therapy completion ([Fig. 1]). Of these, 336 (70.6%) received nitrofurantoin or SMZ-TMP and 140 (29.4%) received
nFLT. The most common nFLT used was ampicillin (13%), followed by cephalexin (10%).
Sixty-eight percent of women received nitrofurantoin; this was the most common treatment
and the most common FLT ([Table 1]).
Fig. 1 Study population.
Table 1
Antibiotics received among women with UTI and ASB
|
Antibiotic
|
N (%)
|
|
First-line therapy
|
335 (71)
|
|
Nitrofurantoin
|
322 (68)
|
|
SMZ-TMP
|
13 (3)
|
|
Nonfirst-line therapy
|
140 (29)
|
|
Ampicillin
|
60 (13)
|
|
Cephalexin
|
49 (10)
|
|
Other
|
31 (6)
|
Abbreviations: ASB, asymptomatic bacteriuria; SMZ-TMP, trimethoprim-sulfamethoxazole;
UTI, urinary tract infection.
Women who received FLT were more likely to have a BMI > 40 (p = 0.04) than those who received nFLT ([Table 2]). Other baseline demographic data did not differ ([Table 2]). Women who were treated with nFLT were more likely to receive a short course of
their prescribed antibiotic ([Table 2]).
Table 2
Maternal demographics among women with ASB and UTI by antibiotic treatment regimen[a]
|
Non-1st line Tx
n = 140 (%)
|
1st line Tx
n = 335 (%)
|
p-Value
|
|
Median age, years (IQR)
|
27 (22, 32)
|
28 (23, 33)
|
0.34
|
|
African American Race
|
31 (22.1)
|
83 (24.7)
|
0.64
|
|
Private insurance
|
43 (31.4)
|
116 (35.1)
|
0.15
|
|
Limited prenatal care
|
7 (5.1)
|
14 (4.2)
|
0.81
|
|
Chronic hypertension
|
16 (11.5)
|
23 (6.9)
|
0.10
|
|
Diabetes mellitus
|
9 (6.5)
|
17 (5.1)
|
0.52
|
|
Tobacco use
|
14 (10.1)
|
36 (10.8)
|
>0.99
|
|
BMI ≥40
|
29 (20.7)
|
43 (12.8)
|
0.04
|
|
Median baseline hematocrit (IQR)
|
36.6 (35.0,39.9)
|
37.0 (34.6,39.0)
|
0.49
|
|
Depression
|
15 (10.7)
|
68 (20.2)
|
0.01
|
|
Multiparous
|
81 (57.9)
|
213 (63.4)
|
0.30
|
|
Prior preterm birth
|
24 (17.1)
|
52 (15.5)
|
0.68
|
|
History of pyelonephritis
|
4 (2.9)
|
14 (4.2)
|
0.61
|
|
History of UTI
|
27 (19.3)
|
90 (26.9)
|
0.10
|
|
Median GA of first infection, wk, (IQR)
|
11.0 (8.7,21.3)
|
11.6 (9.3,18.4)
|
0.54
|
|
Short therapy (<7 d)
|
16 (11.4)
|
12 (3.6)
|
<0.01
|
Abbreviations: ASB, asymptomatic bacteriuria; BMI, body mass index; GA, gestational
age; IQR, interquartile range; Tx, therapy; UTI, urinary tract infection.
a Data presented as n(%), unless otherwise stated.
E. coli was the single most common pathogen, occurring in 41.8% (n = 199) of all positive cultures. Among women receiving FLT, 48.5% had a culture positive
for E. coli, compared with only 25.7% of women who received nFLT, (p < 0.01). Group B Strep LUTI were also common, occurring in 31.7% (n = 151) of all cultures. Women with GBS LUTI were less likely to receive FLT, (24.1%
FLT vs. 50.0% nFLT, p < 0.01).
Among the 476 women included in the analysis, 35 (7.3%) developed pyelonephritis.
Progression to pyelonephritis did not differ by receipt of FLT, (8.2 vs. 5.8%, p = 0.44). After controlling for history of UTI, culture result, treatment duration
<7 days, and BMI ≥ 40, no difference in odds of progression to pyelonephritis was
seen between women who received FLT and those that did not (adjusted odds ratio 1.02,
95% confidence interval 0.42, 2.49). Pyelonephritis-associated morbidities were uncommon.
However, of those women who progressed to pyelonephritis in the nFLT and FLT groups,
a trend toward increased rates of sepsis was seen in those who received nFLT (2/8
vs. 0, p = 0.05). Regarding obstetric outcomes, receipt of FLT was also not associated with
earlier gestational age at delivery, PTB, or LBW ([Table 3]), even after controlling for confounders (data not shown).
Table 3
Maternal and obstetric outcomes among women with ASB and UTI by antibiotic treatment
regimen[a]
|
Non-1st line Tx
n = 140 (%)
|
1st line Tx
n = 336 (%)
|
p-Value
|
|
Progression to pyelonephritis
|
8 (5.8)
|
27 (8.2)
|
0.44
|
|
Median pyelonephritis LOS, days, (IQR)
|
3.5 (2, 5)
|
4 (2, 5)
|
0.97
|
|
Median hematocrit nadir (IQR)
|
26.7 (24.4, 29.9)
|
29.9 (27.0, 31.8)
|
0.12
|
|
Anemia
|
7/9 (77.8)
|
21/27 (77.8)
|
0.99
|
|
Blood transfusion
|
1/8 (12.5)
|
0
|
0.23
|
|
Sepsis
|
2/8 (25.0)
|
0
|
0.05
|
|
Median GA delivery, weeks, (IQR)
|
39.3 (37.7, 40.0)
|
39.1 (37.9, 40.0)
|
0.94
|
|
Preterm birth
|
13 (9.3)
|
42 (12.5)
|
0.35
|
|
Low birth weight
|
15 (10.7)
|
23 (6.9)
|
0.19
|
Abbreviations: ASB, asymptomatic bacteriuria; GA, gestational age; IQR, interquartile
range; LOS, length of stay; Tx, therapy; UTI, urinary tract infection.
a Data presented as n(%), unless otherwise stated.
Comment
Among women with LUTI, receipt of nonfirst-line antibiotic therapy was not associated
with higher odds of progression to pyelonephritis. Pyelonephritis-related sepsis was
uncommon in both groups; however, it occurred in 25% of women in the nFLT compared
with no women in the FLT group. No other pyelonephritis-related morbidities differed
by antibiotic receipt, nor did obstetric outcomes.
Treatment of LUTI in pregnancy is restricted due to antibiotic safety and toxicity.
Outside of pregnancy, where antibiotic selection is not limited by pregnancy-specific
safety concerns, the recommended therapies for UTI in women include fosfomycin, nitrofurantoin,
and SMZ-TMP.[12] Unlike β-lactams, these first-line agents in nonpregnant women have less microbial
resistance and achieve higher rates of complete clearance of gram-negative microorganisms
from the urine, decreasing LUTI recurrence rates.[13] In pregnancy, guidelines identify nitrofurantoin and SMZ-TMP and first-line treatment
of LUTI in pregnancy based on efficacy and teratogenicity.[10] A large recent observational study of more than 450,000 pregnancies found that only
42.3% of first trimester LUTI are treated with FLTs of nitrofurantoin or SMZ-TMP.[8] In the current study, 70.6% of women were treated with first-line therapies suggesting
that national rates of first-line therapy may vary by institution and provider preferences.
A 2011 Cochrane review of 1,125 women treated for AC in pregnancy found no significant
differences between antibiotic treatments with regards to infection cure rate, but
did not assess progression to pyelonephritis.[14] This study also noted that it was not possible to draw conclusion on the best class
of antibiotic nor the duration, and went on to say that “Future studies should evaluate
the more promising classes of antibiotics, such as nitrofurantoin, trimethoprim-sulfamethoxazole…”[14] Thus, more data are needed. However, the findings of our study appear consistent
with these results and are more current; given that existing data suggest that cure
rates are similar, progression rates also likely do not differ.
Use of nFLT for treatment of LUTI in pregnancy does not appear to carry increased
risk of progression to pyelonephritis or worse obstetric outcomes. These findings
suggest that antibiotic choice for LUTI in pregnancy may not need to be limited to
the recommended first-line therapies. Thus, the choice of antibiotics for treatment
of LUTI can be based on other factors including provider or patient preference, local
resistance patterns, and side effect profiles. Further larger studies are needed to
assess rates of progression to pyelonephritis among individual antibiotics within
the FLT and nFLT groups. Additionally, prospective randomized control trials could
provide more data on ideal FLT in pregnancy.
This study has multiple strengths. First, the study was conducted over 6 years at
a large health care system allowing for a representative sample of antibiotic therapies
used for LUTI in pregnancy. Additionally, the study population was diverse across
metrics of age, race, ethnicity, and socioeconomic status making it generalizable
to the broader population. We also recognize several limitations of the study. In
this study, antibiotic therapies were simplified into only two groups: FLT and nFLT.
Individual antibiotics within the respective groups may not share the same outcomes
as others in the group. Given the limited sample size of this cohort, we were only
able to analyze between general therapy groups. The sample size thus limits the ability
to differentiate among therapies and provide detailed outcomes for each individual
therapy. We also recognize that among gram-negative infections in this cohort, there
is some degree of antibiotic resistance; however, in the adjusted analysis, we controlled
for infectious pathogen to mitigate this to some extent. Another limitation concerns
the rate that a test of cure was performed within 2 weeks of treatment, only 12% of
the time, which limits our ability to comment on effectivity of treatment. However,
a larger analysis could be completed to evaluate outcomes of each antibiotic independently.
We are also limited by a small sample size. With the current sample size, a significance
level of 0.05, and assuming an increase from 5 to 10% incidence of pyelonephritis
as a clinically important difference to detect, the power of this study to exclude
a type II error is 45%. Achieving 80% power and assuming the same clinically important
difference to detect would require 474 subjects per group. Thus, there is a chance
of type 2 error. Additionally, very few women, only 12%, had a test of cure within
2 weeks of completion of therapy, limiting our ability to comment on bacterial clearance.
We are also unable to explain antibiotic prescribing that varied from our standing
protocols which recommend nitrofurantoin, cephalexin, or SMZ-TMP, as this occurred
at the discretion of the provider (of which there are over 100 involved in the care
of pregnant women at our institution). Lastly, the rate of progression to pyelonephritis
in this study was found to be 7.3% whereas prior large analyses demonstrate the rate
of 1.4%.[15] This could suggest that our population of patients was at higher risk of pyelonephritis
than seen nationally regardless of the treatments used, making the results less generalizable.
LUTIs in pregnancy treated with nonfirst-line antibiotics do not appear to progress
to pyelonephritis at higher rates than those treated with nitrofurantoin or SMZ-TMP.
Additionally, no difference was seen in obstetric outcomes in women treated with first-line
antibiotics versus nonfirst-line antibiotics. Future studies are required to determine
whether individual antibiotics used in pregnancy, regardless of first-line or nonfirst-line
status, result in higher rates of adverse obstetric outcomes or rates of progression
to pyelonephritis.
