The Rise of Fluoroquinolone-Resistant Escherichia coli in the Community: Scarier Than We Thought

(See the major article by Gurnee et al on pages 1862–8.)

Physicians have long had reliable antibiotics for dealing with common community-associated infections caused by Enterobacteriaceae, including Escherichia coli. However, we have had no new classes of antibiotics to treat resistant gram-negative bacilli since the quinolones, a class of antibiotics which is now 50 years old. Unfortunately, the microbes we deal with have not accommodated our lack of innovation by slowing their spread of antibiotic resistance.

Rates of quinolone resistance among Enterobacteriaceae in the community remained low for decades. In the mid-to-late 1990s, resistance rates were <1% [1, 2], and they rose only to 1%–3% as late as 2008 [3, 4]. However, quinolone resistance rates have skyrocketed in recent years, such that now >10%–30% of community-associated Enterobacteriaceae are quinolone resistant in many parts of the United States [5–8], and much higher rates (>50%) are seen in other parts of the world [9–11]. This dramatic increase in quinolone resistance has been closely associated with spread of a specific clone of E. coli, ST131-H30 [5, 12, 13].

The rise in resistance has necessitated a sea change in medical practice. No longer can doctors “fire and forget” quinolones at community-associated urinary tract, abdominal, or prostatic infections. Cultures and close follow-up are now required. Indeed, quinolones are no longer considered first-line options for treating cystitis [14], and, unfortunately, we have little to replace them with for the treatment of pyelonephritis, which continues to be a major cause of morbidity in otherwise healthy young women [15]. Bringing such patients back to the hospital for intravenous therapy has become common because of inability to treat with any oral agents. How did we get to this point?

It is rational to hypothesize that direct exposure of an individual to antibiotics selects for resistant bacteria. Gurnee et al, in this issue of The Journal of Infectious Diseases, have challenged this notion with an alarming and informative exploration into the transmission dynamics of ciprofloxacin-resistant E. coli (CREC) in children and their mothers [16]. The investigators took advantage of a healthy twin cohort in St. Louis to examine the colonization transmission of CREC from mothers to children in the presence or absence of antibiotic pressure.

Women with twin pregnancies were longitudinally monitored prior to birth, and then they and their twin children were followed postpartum for a median duration of approximately 2.5 years, with serial stool samples. It is alarming that one third of the households of these healthy mothers and children had at least 1 stool specimen with CREC, particularly considering the healthy nature of the population studied. Overall, approximately 20% of children or mothers had at least 1 stool specimen with CREC. Since the lower limit of detection was 141 CREC per gram of stool—as determined by spiking experiments—and people produce tens to hundreds of grams of stool per day, the numbers of positive stool specimens in the study are likely an underestimation of the true burden of colonization. Extrapolating these results, the implications are frightening regarding the number of people in the United States (20% [65 million of 320 million]) and on Earth (20% [1.4 billion of 7 billion]) who may periodically harbor CREC in their intestines.

Children who had a stool specimen positive for CREC were more likely to have had a longer hospital stay after birth and were more likely to have a mother with a positive stool specimen. Surprisingly, no other factors were associated with having a CREC-positive stool specimen, including receipt of antibiotics by either the mothers or the children. Indeed, 6 children had a CREC-positive stool specimen despite having never received antibiotics, and another 2 had positive stool specimens after having received antibiotics >9 months earlier. Nearly half the mothers and children with a CREC-positive stool specimen did so repeatedly and consecutively, indicating that colonization can persist at levels above the threshold of detection for some time, even in the absence of ongoing selective pressure. In 6 of 9 families in which children had CREC-positive stool specimens, both twins were affected and typically had multiple positive samples. These results suggest that persistent colonization may result from a shared environmental exposure or perhaps even from person-to-person spread within the household, which previously was demonstrated to occur with resistant E. coli [17].

The vast majority of the CREC strains belonged to ST131-H30. Such strains are almost always quinolone-resistant, are the predominant cause of extended-spectrum β-lactamase–producing E. coli infections in humans in the United States and globally and have been described as being hypervirulent, based on worse clinical outcomes as compared to less resistant strains [5, 12, 13]. Thus, a frighteningly high percentage of otherwise healthy mothers and children carry multidrug-resistant E. coli in their intestines, at least sporadically and even in the absence of taking antibiotics.

The primary limitation of the study is the small sample size and geographical restriction of the selected population. How generalizable these data are to other populations is not entirely clear. However, these were healthy women and newborns. Furthermore, the assay to detect the presence of CREC was relatively insensitive and likely missed colonization that was below the limit of detection. Thus it is likely that these results underestimate the scope of carriage of CREC in the United States and internationally.

In summary, carriage of potentially pathogenic and virulent, highly antibiotic-resistant E. coli occurs with alarming frequency among healthy women and children, including those who have not knowingly been exposed to antibiotics. One potential explanation for the frequency of carriage despite not receiving antibiotics is spread by contact with or via fomite intermediates from other people who have received antibiotics. Another potential explanation is the inadvertent consumption of antibiotics or antibiotic-resistant bacteria from the food supply, whether meat, dairy, or vegetables [18–20]. Indeed, quinolones have been heavily used in agriculture and aquaculture for many years [20].

We put approximately 15 million kg of antibiotics into the environment every year in the United States alone [19]. Eighty percent of that antibiotic use is in food production. Reduction of inappropriate use of antibiotics in both human and agricultural settings is necessary. Otherwise, the frequency of carriage of resistant bacteria in and on our bodies will continue, and the morbidity and mortality of resistant infections will continue to increase.

These results strongly underscore the interconnectedness of our world. Antibiotics are a shared societal trust, and the spread of antibiotic resistance reflects the outcome of our shared violation of that trust. We need to work together to counteract, and thereby solve, the problem.