In this study, the urinary exposure to nitrofurantoin was investigated using the two common dose regimens in women over 18 years of age with complaints of an uncomplicated UTI. There was no significant difference in exposure in the urine between the two regimens. The difference in exposure was not statistically significant between these regimens for age, renal function, and clinical outcome. The outcome was not related to the urine nitrofurantoin exposure nor to the regimen, age, or renal function.

There are not enough consistent data in the literature to establish a PK/PD index for nitrofurantoin. For Enterobacter cloacae, a concentration-dependent bactericidal effect was demonstrated, while for E. coli and Klebsiella pneumoniae, the effect was time dependent [27]. Since the PK/PD index and the minimally required exposure are unknown, the clinical EUCAST (European Committee on Antimicrobial Susceptibility Testing) breakpoint is based on the ECOFF (epidemiological cutoff value) and used in uncomplicated UTI due to E. coli only [28]. Here, we could demonstrate no difference in urine exposure to nitrofurantoin in terms of AUC for the two most commonly used dosages. In our study, the PK of nitrofurantoin was independent of the dose and formulation of the dose regimens. These results suggest that the urinary PK of nitrofurantoin is not dependent on dose or formulation alone as was described by ten Doesschate [29]. Possible other contributing factors of PK could be lifestyle, fluid intake, nutrition, and voiding pattern.

We found a high variability in AUC0-24h values in our patients, but it was independent of the formulation and dose. A high variability of PK values was described in several studies in the literature [28,29,30,31]. The variation could be explained by differences of included participants (patients or healthy volunteers), dose regimens, or formulations and by differences in renal function, urine frequency, and drugs adsorption. Variation in PK is a common phenomenon for drugs used in the treatment of UTI and not dependent on dose and formulation alone in healthy volunteer studies with no defined time for urine sampling [30,31,32]. Another possible explanation for the high variability in PK studies reporting urinary exposure is the difference in fluid intake between patients. We did not record the fluid intake to mimic real-life circumstances. The results are in line with the study of Huttner et al. [33]; we did not find differences in urinary exposure between the 50 mg and 100 mg doses, although the authors found a twice as high plasma concentrations with the dose regimen of 50 mg compared with 100 mg. The great variability in number of samples received might explain in part the variation in AUC values in different studies.

In clinical studies, some authors [34] have suggested that the slow dose release (100 mg q12) has higher cure rates than the immediate release capsules, but this is not supported by our data concerning the exposure to nitrofurantoin in the urine.

Two to four weeks after finalizing the course, five patients (13%) went back to the GP as they had persisting or recurrent UTI complaints. This cure rate of 87% was in line with the 70% found in a randomized clinical trial using 100 mg q8 as dose regimen in a comparable patient population [35]. The variation in cure rate might be due to the large individual variability in urinary concentrations resulting in sub-therapeutic concentrations in some patients or the limited number of patients included in the study. The 87% cure rate in our study was based on the number of patients (n = 5) returned to the GP with UTI complaints and of the results of the telephone interviews of all patients. No additional treatment failures were mentioned by telephone. Therefore, we considered the 87% cure rate as correct, since patients have their own GP and close communication of emergency rooms to their own GP; no anonymous walk in clinics is present in the Netherlands.

Our study has some strengths and limitations. It is the first study comparing AUC data of two commonly used dose regimens of nitrofurantoin with respect to age, eGFR, urinary exposure, and clinical outcome. As the use of nitrofurantoin was discouraged in the elderly and in patients with impaired renal function, our intended patient population focused on this kind of patients. Since it was an observational study, this limited the inclusion in the study. Other limitations are that we included only female patients as most patients visiting general practitioners with complaints of an uncomplicated urinary tract infection are female patients. Because of common practice, these patients are treated without bacteriological culture of the urine samples. The lack of culture results is another limitation which might influence the cure rate. However, even with these small number of patients, no difference of the urine exposure to nitrofurantoin was found. Furthermore, at the beginning of the study, we wanted to include patients with a wide range of eGFR (starting from 30 mL/min). As the study had a naturalistic design, performed at GPs, we were only able to include women with eGFRs between 50–59, 60–89, and ≥ 90 mL/min, as the GPs were reluctant to prescribe nitrofurantoin to women with an eGFR < 50 mL/min. Therefore, no conclusion can be drawn on the effect of a reduced eGFR (below 50 mL/min) on the urinary exposure of nitrofurantoin.

In conclusion, based on the data of our study, we found that a dose of 50 mg q6h or 100 mg q12 nitrofurantoin was comparable and that there is no preference of one above the other dose regimes. To answer the question as to the applicability of nitrofurantoin in the elderly and in those with impaired renal function, studies with more patients are needed.