Introduction

A growing body of literature has associated nocturia with significant morbidity and mortality. Nocturnal voiding is a cardinal feature of several hypervolemic conditions, and may be the presenting sign of serious systemic cardiovascular, respiratory, endocrine, and metabolic diseases [1]. Although the cause-and-effect relationship between nocturia and mortality remains to be established, the multitude of comorbid conditions and their association with nocturia risk factors aligns well with the accumulated deficits model of frailty, which considers frailty as a function of the total burden of health deficits acquired across the aging process [2, 3]. Frailty has been conceptualized and measured using a variety of approaches [4,5,6,7,8], but one of the most widely used is the frailty index (FI), a valid and reliable measure derived from the Canadian Study of Health and Aging (CSHA) [2,3,4]. The FI captures the presence of current diseases, performance in activities of daily living, and physical signs from the physical examination, and higher levels of the FI are closely correlated with increased morbidity and early mortality [2,3,4].

The pathogenesis of nocturia relies upon a fundamental mismatch between nocturnal urine volume and storage, wherein nocturnal voiding may be driven by excess production, as seen in nocturnal polyuria and global polyuria, and/or reduced bladder capacity (either functional or anatomic) [1]. However, to our knowledge, an association between nocturia and frailty has never been described. Accordingly, we sought to use voiding diary data from elderly men with LUTS at an outpatient urology clinic to characterize the relationship between nocturia and frailty.

Methods

Study design and procedures

This study was a retrospective analysis of 24-h voiding diaries completed from 2008 to 2018 by elderly male veterans who were being treated for lower urinary tract symptoms (LUTS) at a Veterans Affairs outpatient urology clinic.

All patients had established care before being asked to complete a voiding diary. Patients were being actively managed by a urologist in accordance with the best practice framework for the evaluation and treatment of nocturia [9], which involved a thorough medical interview and physical examination, with individualized behavioral modification plans (evening fluid restriction, bladder training exercises, etc.) as a first-line intervention; medications for persistent symptoms; and outlet-reducing surgery for urinary retention refractory to pharmacologic therapy [9].

A voiding diary database for retrospective analysis was compiled with approval from the Veterans Affairs New York Harbor Healthcare System Institutional Review Board. A waiver of informed consent was granted as voiding diaries are a standard of care in the evaluation and management of LUTS [9]. Patient electronic records were reviewed to determine demographics, medications, genitourinary diagnoses, procedures, and comorbidities.

Patients were included if they were male, age ≥ 65 years, and had completed at least one 24-h voiding diary showing ≥ 2 nocturnal voids during the sleeping period. Only the first diary showing ≥ 2 nocturnal voids was included from patients with more than 1 complete diary. Patients were excluded if they had a diagnosis of lithium-induced diabetes insipidus because of the pronounced association between lithium-containing drugs and polyuria (n = 4) [10].

Data from voiding diaries were collected in a standard manner (as modified from van Kerrebroeck and colleagues) [11] and analyzed retrospectively (Table 1). Although the FI, as originally described, is comprised of 92 items [2], individual studies often use subsets or smaller numbers of items with which to compute the FI [6]. In our study, the LUTS database included 39 of the 70 possible clinical assessment items (out of the total of 92 items) (see Table 2 for a listing of those included in the current study). As is customary for studies using subsets of items from the original CSHA [5, 6], we calculated the FI on the basis of the proportion of conditions in any given patient relative to the total number of conditions in the item pool. This approach was also used in the original description of the FI [2].

Table 1 Overview of voiding diary parameters
Full size table
Table 2 Variables that comprise the Canadian Study of Health and Aging (CSHA) 70-item frailty index clinical assessment
Full size table

Many studies have employed FI as a categorical variable and defined frailty using FI thresholds, which often differed depending on the population under study [2,3,4, 12]. The distribution of FI scores in our sample differed from the gamma shaped distribution typically reported in those studies and instead clearly represented a biomodal distribution (Fig. 1). In view of this, we divided our sample into groups representing low, intermediate, and high levels of frailty, based on corresponding FI cut points of ≤ 0.077, {> 0.077 and < 0.179}, and ≥ 0.179, respectively.

Fig. 1
figure 1

Frequency Distribution of modified frailty index. Note X-axis represents frailty index (FI) intervals. To approximate tertiles, low, medium, and high Frailty groups were defined by FI values of 0 to ≤ 0.077; 0.078 to 0.179; and ≥ 0.179, respectively

Full size image

Statistical analysis

For both patient demographics and diary parameters, categorical variables are reported as frequency (percentage), and continuous variables are reported as median (95% confidence intervals) using Wilcoxon confidence interval estimates. Categorical and continuous variables were compared using the Fisher’s exact test and the Kruskal–Wallis test, respectively. When inter-group differences were found to be significant, the Wilcoxon rank-sum test was used to determine partial order between sample pairs. All tests performed were two-sided, and a p value < 0.05 was deemed statistically significant, with the Bonferroni correction applied for multiple comparisons.

Results

The frequency distribution of the FI for the patients generating the 158 patient diaries meeting inclusion criteria is shown in Fig. 1. FI cutoffs were {≤ 0.077} for low (n = 59), {> 0.077 and < 0.179} for intermediate (n = 58), and {≥ 0.179} for high (n = 41) for the predefined groups.

A complete overview of patient demographics, comorbid conditions, and urologic treatment history is provided in Table 3. There were no statistically significant differences in demographics, medication usage, or history of surgery or radiation therapy between frailty groups. Among medical comorbidities, diabetes mellitus (DM), chronic obstructive pulmonary disease (COPD), obstructive sleep apnea (OSA), and congestive heart failure (CHF) were more likely in the high frailty group, with DM occurring in over 25% of the sample overall (44/158).

Table 3 Demographics, comorbid conditions, and treatment history by frailty burden
Full size table

Table 4 indicates the rates of several measures of nocturnal urine production (NUV, MVV, NMVV, and NUP) in the high frailty group compared to the intermediate and low frailty groups. Bonferroni-adjusted pairwise comparisons ([p = 0.05]/3 = 0.017) reached significance for many of these. Additional analysis limited to those cases without DM (n = 114) showed that high frailty continued to be significantly associated with greater nocturnal urine production (NUV, NMVV, NUP; all p < 0.02), but without significant differences in 24-h urine production (p = 0.18).

Table 4 Voiding diary parameters by frailty burden
Full size table

Discussion

In this group of aged male veterans, higher frailty was associated with greater nocturnal urine production. Although some of this effect might have reflected the disproportionate number of DM patients in the high frailty group (as implied by their higher rate of 24-h urine production, consistent with possible glycosuria), stratification analyses showed that the association between nocturia and frailty persisted independently of DM.

Relationships between frailty and DM in old age are well-acknowledged in the literature [13, 14] and the presence of DM has been shown in several studies to be associated with incident frailty [15, 16] or absence of improvement in frailty status [17]. Although the specific mechanisms of how diabetes may affect frailty remain uncertain, likely candidates may include chronic inflammation and/or the adverse impact of skeletal muscle damage and neuropathic pain on mobility and ambulation [18]. Higher rates of 24-h urine production were also seen in the much smaller number of CHF patients, a group with an increased non-osmotic drive for thirst and high utilization of diuretic agents [19, 20], as well as COPD and OSA, but owing to the relatively low prevalence of these conditions in our sample, we were unable to make any valid statistical inferences about the influence of CHF, COPD, or OSA in this data set.

Although examination of the relationship between nocturia and frailty is novel, frailty has been implicated in other LUTS. Frailty is associated with an increased risk of incontinence [21]. Frail older people, as assessed by an impaired timed up and go test, are also more likely to have overactive bladder (OAB) than age-matched non-frail older people [22]. In addition, among community-dwelling older males, severe symptoms on the International Prostate Symptom Score (IPSS) are associated with a higher prevalence of frailty [23]. In the setting of OAB, benign prostatic hyperplasia (BPH), and other LUTS, a reduction in functional bladder capacity is central to the pathogenesis of nocturia in some patients, and conversely, nocturia may be classified as one component of the symptom complex for patients with OAB, BPH with bladder outlet obstruction, or urinary incontinence [1]. However, nocturia is a complex and multifactorial condition, and comorbid LUTS are far from the only underlying etiology.

Unfortunately, older people—and particularly the frail elderly—are often either overtly or covertly excluded from involvement in research studies [24], which has left them underrepresented in pivotal trials on pharmacologic and surgical management for LUTS [25]. Moreover, among older people with LUTS, those with significant frailty may be disproportionately affected by the presence of urologic symptoms as urinary incontinence is associated with an increased risk of hospitalizations, nursing home admissions, and social isolation [26]. Among patients with refractory idiopathic detrusor overactivity, frail older patients experienced a significantly longer duration of recovery and lower long-term success rate following onabotulinumtoxin A treatment compared to their non-frail contemporaries [27]. Among individuals undergoing transurethral resection of the prostate (TURP), frailty is associated with an increased occurrence of complications and discharge to a skilled or assisted living facility, which persists after adjusting for factors including age, race, and recent weight loss [28, 29].

The causes and underlying mechanisms of nocturia among frail elderly men and women are multifactorial and, amongst others, include conditions such as peripheral edema, elevated natriuretic peptide secretion (possibly because of sleep apnea), excessive fluid intake, circadian blunting in arginine vasopressin secretion, medications, or renal tubular dysfunction [23, 30, 31]. Desmopressin, which is used in patients with nocturia owing to nocturnal polyuria and no identifiable contributory comorbidities, is not recommended in frail elderly patients due to an increased risk of hyponatremia [32].

This study is subject to a number of limitations. First, our study represents a retrospective single institution cross-sectional analysis based on clinical records. Although a reliable FI which predicts outcomes can be constructed from as few as 20 [33] or even 11 items [34], our frailty model admittedly could not completely duplicate that used in the CSHA. Given the various approaches to characterizing frailty [5], our use of the FI was limited to the specific conditions accessible to us in our patients’ Veterans Affairs records. Specifically, the frailty model used here could not take into account impairment in daily activities or cognitive impairment, which are both important facets of frailty, such that information regarding actual impairment is limited (except for bladder function and cognitive problems). As the FI reflects mainly medical comorbidities, and patients in the high FI group were disproportionately affected by a number of conditions known to cause increased urine production (i.e., DM, COPD, OSA, and CHF), the results of the present study may be a function of comorbidity burden. Although no inter-group differences were observed in the prevalence of comorbid genitourinary diagnoses, the association between frailty and other LUTS such as OAB and urinary urge incontinence may nevertheless pose an operative systematic bias. Given that only elderly male veterans were included in this analysis, results may not be generalizable to all elderly men due to differences in lifestyle and distribution of associated comorbid conditions. Likewise, these results cannot be extrapolated to women. Despite these limitations, to our knowledge, the present analysis is the only study which has applied a validated method of calculating an FI and explored its relationship with nocturia.

Although the predefined cut points allowed for a comparison of voiding diary parameters between 3 groups of approximately equal size, our cutoff for “high frailty” was significantly lower than other meaningful FI cutoffs that have been identified in population studies [3, 4, 35]. These relatively lower values may reflect the fact that our sample consisted of what some have termed a “young-old” group in their late 1960s and early 1970s, unlike the studies of Rockwood et al. [4] and Hoover et al. [35], which included sizeable numbers of persons in their 1980s and above. Future prospective research involving larger well-defined frailty subgroups is needed to further phenotype the etiology of nocturia in frail individuals and establish individualized recommendations for evaluating and managing nocturia in these patients.

Conclusion

A voiding diary analysis of elderly males with nocturia at an outpatient urology clinic identified frailty as a condition significantly associated with increased nocturnal urine production. Future research on the mechanistic relationship between urine production and functional impairment is warranted.