Sleep is a crucial factor in the maintaining one’s overall health and well-being. Both insomnia and nocturia are known disruptors of physiological sleep patterns and are linked to several pathophysiological responses such as hypertension, cardiovascular disease, hormonal disorders and an overall decreased quality of life. Even though both insomnia and nocturia are prevalent, highly comorbid conditions and share a mutual connection, research considering both conditions simultaneously  remains scarce. The objective of this retrospective analysis was to construct a predictive model for nocturia, using retrospective clinical and polysomnographic data from the TUCSON study: Tackling Underlying Causes Of Sleep Related Nocturia. We hypothesized that certain variables would be strong cross-sectional predictors of nocturia in patients with insomnia and could thus be used to inform a predictive model, enhancing the clinical assessment and management of nocturia.

A retrospective dataset from the TUCSON study was created including all adult patients consulting the Center for Integrative Medicine for insomnia complaints between 2019 and 2021. Patients who underwent a polysomnography (PSG) and filled in the questionnaire were eligible to participate in the study. Patients were excluded from the dataset if the total sleep time was too short (sleep efficiency <50%) in order to accurately calculate the different sleep stages. Data on demographics, medical history, medication use, self-reported nocturia frequency, insomnia severity index (ISI), Pittsburgh Sleep Quality Index (PSQI), polysomnographic sleep parameters and First Uninterrupted Sleep Period (FUSP) were collected. Patients waking up ≥2 to void/night were assigned to the nocturia group. The duration in minutes from “lights out” to the first wake episode of 3 minutes on the PSG was used as a surrogate marker for FUSP. Descriptive statistics and subsequent bivariate analysis of variables related to nocturia were performed in order to identify factors with a p-value less than 0.25, which were deemed potentially significant. These variables were incorporated into a Classification and Regression Trees (CART) model to predict nocturia using the R 'rpart' package. The model was trained on 70% of the dataset and tested on the remaining 30%. The model's performance was assessed using accuracy, sensitivity, specificity, and the area under the Receiver Operating Characteristic (ROC) curve.

The retrospective dataset analyzed comprised a total of 170 patients presenting with insomnia complaints, illustrated by a median ISI score of 20 (IQR 16 – 20), of which 106 patients without nocturia and 64 patients with nocturia (Table 1).  The median age of the total study group was 45 (IQR 31 – 55)  years old of which 58.2% were women. Following bivariate analysis, significant factors for CART model included older age (p=0.185), presence of cardiovascular disease (p=0.127), higher caffeine consumption(p=0.206), lower sleep efficiency (p=0.198), REM as percentage of total sleep time (p=0.113), higher Apnea-Hypopnea Index (AHI) (p=0.058), and shorter FUSP (p<0.001). The CART analysis (Figure 1) determined that the FUSP was the most significant predictor for nocturia, followed by the AHI and the percentage of REM sleep, in descending order of importance. The predictive model's accuracy was 73.08%, with a sensitivity of 79.41%, specificity of 61.11%, balanced accuracy of 70.26%, and an ROC area of 0.82, underscoring the model's clinical discriminative capability.

To our knowledge, this database is the first in its kind to look into close detail to the cross-sectional predictors of nocturia in such a specific insomnia population with the availability of polysomnographic data. This model clearly shows that FUSP is the most important factor in predicting nocturia in patients with insomnia. CART models provide the advantage of developing prediction models for which regular regression models lack a good fit. Moreover, the AUC of 0.82 indicates FUSP is not only the most important factor but also has a good discriminatory power. This is an interesting finding as a shorter FUSP has been associated with lower sleep quality and duration, resulting in daytime complaints. [1] Only recently, the effectiveness of FUSP has been evaluated as a predictor of therapeutic outcomes in children with nocturnal enuresis. [2] This is of clinical interest as patients with insomnia and a short FUSP could potentially benefit from a urological consult in order to treat nocturia and improve their quality of sleep. 
This study has limitations. First this convenience sample comprises a small dataset in a very specific population. All the patients included suffered from severe insomnia, as illustrated by very high ISI scores, for which they consulted or were referred to a tertiary care hospital. All patients received a PSG to exclude other causes potentially interfering with their disturbed sleep, which means that some patients were diagnosed with concomitant obstructive sleep apnea or restless legs. Moreover, the dataset included a large age-range of patients, causing potential dilution of the results considering nocturia is a multifactorial condition that can be influenced by age-related matters such as BPO or polypharmacy.  Nevertheless, as these data show, the FUSP forms a clear discriminator between nocturia and no nocturia in patients with insomnia, it is intended to test the model’s external validity in a future prospective cohort. Employing a predictive model for nocturia, despite its seemingly straightforward diagnostic approach through validated questionnaires, holds promising future benefits and applications. From a lifespan perspective, individuals at increased risk for developing nocturia could be detected at an early stage, enabling proactive measures. In terms of monitoring and prognosis, the model could serve to track progression in patients, aiding in the timely adjustment of treatment if deemed necessary.

This study presents the first of its kind data investigating clinical and polysomnographic predictors of nocturia within a select insomnia population, utilizing CART models, enabling more tailored prediction modelling. Our findings underscore the significance of the FUSP as the primary predictor of nocturia in insomnia patients with a robust discriminatory power demonstrated by an AUC value of 0.82. The model will undergo external validation in the future and is a promising first step in improving sleep outcomes in patients with insomnia and nocturia.

Bliwise DL, Holm-Larsen T, Goble S, Nørgaard JP. Short time to first void is associated with lower whole-night sleep quality in nocturia patients. J Clin Sleep Med. 2015; 11: 53-55.
Karamaria S, Dhondt K, Everaert K, Mauel R, Nørgaard JP, Raes A, Van Herzeele C, Verbakel I, Walle JV. First uninterrupted sleep period in children and adolescents with nocturnal enuresis: Added value in diagnosis and follow-up during therapy. Neurourol Urodyn. 2023 Nov 6. doi: 10.1002/nau.25322. Epub ahead of print. PMID: 37929315.

Continence 12S (2024) 101538
DOI: 10.1016/j.cont.2024.101538