It's during residency that most of us develop our enduring practice patterns, the bulk of our clinical knowledge base, and the idiosyncratic points and pearls that make up so much of the daily ED grind. Much of what we learn is borne from an ever-increasing database of randomized trials and vibrant literature that give us guidance in choosing the best therapies for patients.
Beyond much doubt, our shift away from eminence-based medicine (bedside practice dictated by tradition and experience) toward evidence-based medicine has been perhaps one of the greatest leaps forward in modern medicine. In our haste for excellence, however, sometimes we fall far short.
The American College of Obstetricians and Gynecologists (ACOG) released a committee opinion warning in 2011 against using nitrofurantoin (Macrobid) during the first trimester of pregnancy because of the perceived risk of an increased rate of congenital abnormalities with its use. (Obstet Gynecol 2017;130:e150; http://bit.ly/2yc3APG.)
The committee continued to recommend that nitrofurantoin be used as a first-line agent during the second and third trimesters, but stated that it should only be considered appropriate in the first trimester when no other suitable alternative antibiotics were available.
I was shocked. Not only had it been my routine to prescribe nitrofurantoin to pregnant women with urinary tract infections, I had been taught in residency to, in fact, go out of my way to do so. Never before had I seen such a profound and relevant conflict between the worlds of eminence- and evidence-based medicine. This common clinical question had, seemingly, not only been answered incorrectly for so many years, but the ACOG committee opinion suggested we may be causing outright harm with a routine practice.
I searched harder.
The ACOG recommendation would seem to be borne out most significantly from a paper published in 2009. (Arch Pediatr Adolesc Med 2009;163:978; http://bit.ly/2IOpk85.) This case-control study asked mothers by telephone anywhere from six weeks to 24 months following delivery whether they had had a UTI from one month prior to conception until the end of the third month of pregnancy and which antibiotic they were prescribed.
Without any independent verification of the accuracy of patient report, the researchers found that exposure to nitrofurantoin was associated with increased risk of four types of birth defects: anopthalmia or microphthalmos, hypoplastic left heart syndrome, atrial septal defects, and cleft lip with cleft palate.
Oddly, though, the authors found no increased risk of birth defects with tetracycline use (previously classified as pregnancy category D for evidence of human fetal risk), but identified risk of limb deficiencies and atrial septal defects with penicillin and cephalosporin use. Clearly, this single (small) study was flawed, yet it (and another similar but smaller case-control study) led to sweeping recommendations from ACOG.
Years later, larger investigations have offered a more comprehensive and complete understanding of nitrofurantoin's safety in pregnancy. A population-based cohort study in 2013 analyzing more than 180,000 pregnancies found no association between nitrofurantoin and any major malformations, and the rates of septal defects and oral clefts were, in fact, lower in patients exposed to nitrofurantoin than otherwise. (Obstet Gynecol 2013;121[2 Pt 1]:306; http://bit.ly/2EgoUch.)
Further weakening the earlier observations suggesting fetal risk, another large retrospective review was published later that same year examining 1,319 nitrofurantoin exposures among an Israeli cohort, which also found no association with increased risk of congenital malformations. (J Clin Pharmacol 2013;53:991.)
When Data Do Harm
Just as increased knowledge and meticulous investigation can lead us toward better care and outcomes for our patients, incomplete or casually interpreted data can cause true harm. Urinary tract infections complicate up to 10 percent of pregnancies, and untreated and undertreated infections have been associated with low birth weights, preterm labor, and neonatal sepsis. (Am Fam Physician 2000;61:713; http://bit.ly/2IPhknr.)
Cephalexin, recommended first-line in first trimester UTIs by ACOG, is typically effective in the majority of urinary tract infections, but the growing resistance of some isolates to cephalosporins and gram-positive organisms on which cephalexin has no activity such as Enterococcus faecalis (which represented up to 12 percent of cultures in one recent investigation) are concerning. (JAMA 2018;319:1781.)
Most patients with resistant organisms can have their antibiotic regimens adjusted following culture data, but EPs are often faced with patients in whom follow-up and ongoing care are unlikely at best. A significant proportion of these patients are lost to follow-up because of homelessness, socioeconomic disadvantages, and even medical disenfranchisement, prompting a more critical need to get it right the first time.
Certainly, it seems likely that the risk of undertreated infection outweighs the possibility of congenital malformation in patients with recent antibiotic use or in whom testing indicates lack of urinary nitrites (suggestive of an increased likelihood of E. faecalis infection). (J Emerg Med 1997;15:435; J Emerg Med 2010;39:6.)
Nitrofurantoin in pregnant patients with urinary tract infection is a clinical challenge where traditional practice outshines the ACOG committee opinion, which should be revisited in light of the compelling proof of safety and the increasing risk posed by resistant organisms. Emergency physicians should have a low threshold to prescribe nitrofurantoin for patients with urinary tract infections in the first trimester of pregnancy, particularly when poor follow-up is anticipated or there is suspicion for cephalosporin-resistant organisms, that is, patients with recent antibiotic use, previous resistant organisms, or nitrite-negative urinalyses.