Abstract
Background
Hyponatremia has long been recognized as a common electrolyte abnormality among skilled nursing facility (SNF) residents. In the literature, hyponatremia is associated with increases in falls, cognitive impairment, hospitalization, and increased mortality. Although many medical conditions and medications contribute to hyponatremia risk, the role of enteral feeding tubes in this population has not been well defined in clinical practice. This study examined the association between enteral feeding tube use and incident hyponatremia and evaluated selected chronic conditions and medication exposures in SNF patients.
Methods
We conducted a retrospective cohort study using secondary data from a suburban skilled nursing facility. Adults aged ≥ 18 years residing in the SNF with available laboratory data were included (n = 1,256). Exposures included presence of an enteral feeding tube at admission, selected chronic conditions, and high-risk medication exposures. The primary outcome was incident hyponatremia, defined as any corrected serum sodium value < 135 mEq/L occurring after admission. Relative risks were estimated using modified Poisson regression.
Results
Incident hyponatremia occurred in 34.6% of residents. In adjusted models, enteral feeding tubes (aRR 2.42; 95% CI 1.84–3.19), chronic kidney disease (aRR 1.87; 95% CI 1.43–2.44), and Liver Disease (aRR 1.58; 95% CI 1.20–2.07) remained the significant independent predictors. Congestive Heart Failure, ethnicity, sex, age, and high-risk medication exposure were not significantly associated with risk at the predetermined alpha = 0.01, and no significant interactions were observed between CKD, CHF, and liver disease and enteral feeding tubes on incident hyponatremia.
Conclusions
Hyponatremia remains a common electrolyte abnormality among SNF residents. Enteral feeding tube use was strongly associated with increased cumulative risk of incident hyponatremia. Both CKD and liver disease also pose risk of incident hyponatremia. SNF residents with enteral feeding tubes may benefit from targeted monitoring and risk stratification in long-term care settings. Clinicians should be aware of these risks while ordering water flushing of enteral tubes and assess these patients regularly for hyponatremia.
Clinical trial number
Not applicable.
Keywords: Hyponatremia, ; Enteral feeding tube; Geriatrics; Skilled nursing; Skilled nursing facility; Clinical epidemiology; Observational study
Introduction and background
Hyponatremia is often encountered in clinical practice and has a higher prevalence in geriatric clinical practice. A particularly high-risk subgroup for hyponatremia are those living in skilled nursing facilities (SNF) having an existing diagnosis of hyponatremia [1–5]. Prior cross-sectional and longitudinal studies have demonstrated a substantially higher prevalence of hyponatremia among SNF residents compared with community-dwelling older adults, in whom reported prevalence estimates range from approximately 7–11%, while nursing home prevalence may exceed 50% in some studies [6, 7]. Clinical and epidemiological studies consistently link hyponatremia to adverse clinical outcomes, including generalized weakness, falls, fractures, cognitive impairment, hospitalization, and increased mortality [8–13]. Mortality is increased both in-hospital (17.4%) and at 1 year (28.3–41%) compared to eunatremic patients [12, 14]. In patients with the diagnosis of Heart Failure with preserved Ejection Fraction (HFpEF), hyponatremia predicts all-cause mortality, rehospitalization, and stroke [15]. The risk of hyponatremia in this population reflects the intersection of aging, multimorbidity, and care-related exposures, including chronic disease burden and institutional management of fluids and nutrition. Despite its frequency and clinical impact, hyponatremia in SNFs remains under-characterized, particularly with respect to modifiable risk factors and care practices. Few studies have specifically quantified this risk in skilled nursing populations, creating a gap in clinical knowledge. Accordingly, this study aimed to characterize the burden of hyponatremia in skilled nursing facilities and to examine whether enteral feeding tube use and selected chronic diseases are associated with increased risk of hyponatremia in this population.
Review of literature
Hyponatremia in older adults is multifactorial and reflects the interaction of age-related physiologic changes, chronic disease burden, medication exposure, nutritional status, and institutional care factors. Aging is associated with impaired renal free water excretion, altered thirst regulation, and increased sensitivity to antidiuretic hormone, all of which may predispose older adults to sodium dysregulation. Chronic conditions common among skilled nursing facility residents, including congestive heart failure, chronic kidney disease, and liver disease, may further impair water balance through neurohormonal activation and reduced effective circulating volume. Additionally, medications frequently prescribed in older adults, including thiazide diuretics, antidepressants, antiepileptics, antipsychotics, and proton pump inhibitors, have been associated with hyponatremia through mechanisms such as syndrome of inappropriate antidiuretic hormone secretion and impaired urinary dilution.
Incidence and prevalence in SNFs
As previously mentioned, the general estimation of hyponatremia varies by study, but is generally found to be between 15 and 30% with malnourished patients potentially having higher risk [1–5, 12, 13]. In two older studies from Miller [4] (1995) and Kugler & Hustead [2] (2000), up to half of the samples had at least one incident lab value showing hyponatremia in the year before data collection occurred.
Risk factors in the elderly
Several main risk factors were identified in the literature for incident hyponatremia. These factors can be subcategorized into (1) Medication-related factors (e.g., thiazide diuretics, antidepressants, antipsychotics, anticonvulsants, and prolonged proton pump inhibitor use) [1, 2, 16–19], (2) cardiac, renal, and hepatic disease comorbidities [1, 14–16, 19, 20], (3) low solute intake or malnutrition [1, 5, 10, 11, 19], (4) acute illness [1, 8, 19, 21, 22], (5) age-related renal and hormonal changes [1, 2, 16], (6) SIADH and non-osmotic ADH release [1, 4, 19, 22], and (7) institutional care factors [5, 19].
Potential association of hyponatremia and enteral feeding
Some evidence suggests patients receiving enteral nutrition may be at increased risk of hyponatremia due to poor solute intake and high free water administration [1, 2, 5, 16, 19]. One study by Rajarajewwari et al. [23] (2023) reported a negative correlation between the Identification of Seniors at Risk (ISAR) Scoring tool and hyponatremia severity. This may suggest clinical utility of ISAR for risk stratification screening for hyponatremia.
Summary of literature
Despite biologic plausibility and frequent occurrence in the population, relatively few studies have specifically evaluated enteral feeding tube use as an independent predictor of hyponatremia among skilled nursing facility residents.
Methods
Study design and setting
Data was collected and analyzed. Data were collected and analyzed as a quantitative secondary analysis using a retrospective cohort design from a single suburban skilled nursing facility in Illinois that provides long-term care, subacute rehabilitation, ventilator care, and in-house dialysis services. The facility receives both regional and out-of-state referrals and serves a medically complex resident population. Residents admitted between January 2022 and April 2025 were eligible for inclusion. All data were deidentified at the time of collection. This study was reported in accordance with the STROBE guidelines for observational studies. Clinical and laboratory data were extracted from the facility’s electronic health record system, which is routinely updated during resident care by licensed healthcare personnel. Laboratory values were imported directly from affiliated laboratory reporting systems, while diagnoses and medication data were maintained as part of standard clinical documentation and administrative review processes.
Study population and eligibility criteria
Inclusion criteria consisted of current or prior residence at the facility, age ≥ 18 years, and availability of at least annual laboratory data. Exclusion criteria included documented prevalent hyponatremia, defined as serum sodium < 135 mEq/L identified prior to or at the time of index admission or documentation of ICD-10-CM code E87.1 for hyponatremia within the available medical record, pituitary tumors, total parenteral nutrition, diagnosed syndrome of inappropriate antidiuretic hormone secretion (SIADH), hypothyroidism, adrenal insufficiency, or chronic scheduled intravenous fluid administration. Chronic intravenous fluid administration was defined as anyone with a standing order for intravenous fluids given at specified intervals.
History of hyponatremia was determined using available laboratory and clinical records within the electronic health record prior to and at the time of index admission, including outside records uploaded into the resident chart at admission. Participants with any documented serum sodium concentration < 135 mEq/L identified before cohort entry were excluded. Because retrospective laboratory availability varied across residents and outside records were not consistently available, a standardized lifetime look-back period was not feasible.
Exposure and covariates
Enteral feeding tube presence was defined as present or not present on admission and was considered a baseline exposure. Chronic kidney disease (CKD), congestive heart failure (CHF), liver cirrhosis, and high-risk medication use were examined as exposure variables. Demographic variables, including age, sex, race, and ethnicity, were included as covariates. Demographic characteristics and comorbid diagnoses were obtained from admission documentation and ICD-10–coded diagnoses within the electronic health record. Comorbid conditions, including CKD, CHF, and liver disease, were not independently adjudicated using standardized biochemical, imaging, or specialist-confirmed diagnostic criteria; therefore, some degree of diagnostic misclassification remains possible.
Medication exposures were classified based on admission medication status and analyzed as baseline exposures. Participants were classified as exposed if they were prescribed any agent within the following medication classes on admission: antidepressants, antiepileptics, thiazide diuretics, haloperidol, or proton pump inhibitors. For analytic simplicity, these medication classes were compiled into a single composite variable.
Outcome definition
The primary endpoint was cumulative incident hyponatremia, defined as at least one corrected serum sodium concentration < 135 mEq/L occurring after the index admission date. The index admission served as time zero for cohort entry, at which point exposure status was determined. Only sodium measurements obtained after cohort entry were eligible for outcome classification. Sodium values were corrected for hyperglycemia using methods outlined by Hillier et al., (1999) [24, 25]. Serum sodium measurements were obtained using direct ion-selective electrode methodology.
Follow-up duration varied across participants and was determined by the available electronic health record observation period. If any single corrected sodium measurement met the specified threshold during follow-up, the participant was classified as having experienced incident hyponatremia. Repeated hyponatremia episodes occurring in the same participant were not analyzed separately; participants were classified dichotomously based on whether at least one incident episode occurred during the observation period. Serum sodium measurements were obtained as part of routine clinical care rather than through a standardized study protocol. Laboratory monitoring frequency was determined by treating clinicians according to clinical judgment and resident acuity. Because duration of follow-up and laboratory monitoring frequency were not standardized across residents, the outcome was analyzed as cumulative incidence during the available observation period rather than as a time-to-event measure.
Statistical analysis
Baseline group differences between participants with and without enteral feeding tubes and those with and without incident hyponatremia were analyzed using chi-square testing for categorical variables. Age was compared using independent t-tests. Given the retrospective cohort design and binary outcome definition, relative risks (RRs) were selected as the primary measure of association. Because follow-up duration and exact person-time observation could not be reliably quantified across residents, time-to-event modeling and incidence rate estimation per person-year were not feasible; therefore, the outcome was analyzed as cumulative incidence during the available observation period.
Associations between exposures and incident hyponatremia were first evaluated using crude RRs and 95% confidence intervals. Multivariable associations were subsequently assessed using modified Poisson regression with a log link and robust variance estimation (generalized estimating equations) to directly estimate adjusted relative risks. Because the outcome was common (> 10%), modified Poisson regression was selected to avoid overestimation associated with odds ratios from logistic regression. Covariates included demographic and clinical predictors identified a priori. Interaction terms between enteral feeding tube status and selected chronic conditions (CKD, CHF, and liver disease) were included to evaluate potential effect modification. Participants with missing data for the required outcome or baseline predictor data were excluded from the analysis.
Sample size and power calculation
Sample size estimation was conducted for the primary outcome and exposure testing. This was completed using G*Power software based on a 2 × 2 contingency table with one degree of freedom. Given the objectives of multiple comparisons, a conservative two-sided significance threshold of α = 0.01 was applied to primary hypothesis testing to reduce the likelihood of type I error. Statistical power was set at 0.95. Small-to-moderate effect sizes (0.1–0.2) were assumed, the estimated required sample size ranged from approximately 325 to 1,300 participants, depending on the assumed effect size. Based on these assumptions, the study targeted a sample size of approximately 1,300 participants.
Results
Study population and baseline characteristics
A total of 1,349 patient records were available for evaluation. 93 of these records were excluded due to missing data on key exposure or outcome variables, including sodium levels, demographic information, or unclear enteral feeding tube status. The final analytic sample consisted of 1,256 individuals with enteral feeding tubes present in 34.7% of patients, and incident hyponatremia occurred in 34.6% of the cohort. The sample was mainly White (64.3%) and non-Hispanic (85.1%) with a mean age of 66.9 years (SD 14.9; range 24–104). 59.47% of participants were aged ≥ 65 years.
To test for any baseline differences in groups, demographics and clinical risk factors were compared by enteral feeding tube status (Table 1) and by incident hyponatremia status (Table 2). This was done for clinical context rather than investigation of baseline differences. Using a conservative α = 0.01 to maintain consistency in analysis. Both stratification by enteral feeding tube status and incident hyponatremia showed statistically significant baseline differences with race, antidepressant use, haloperidol use, PPI use, CKD status, and CHF status. No statistically significant differences were found between age, ethnicity, sex, composite high-risk drugs, antiepileptics, or thiazides for either stratification by enteral feeding tube status or incident hyponatremia. Additionally, liver disease status differed by enteral feeding tube status but not incident hyponatremia.
Table 1.
Descriptive statistics between groups separated by enteral feeding tube status
| Variable | Enteral Feeding Tube (n = 436) | No Enteral Feeding Tube (n = 820) | p-value |
|---|---|---|---|
| Age, mean (SD) | 65.9 (14.9) | 67.4 (14.8) | 0.08 |
| Race, n (%) | < 0.001** | ||
| Asian | 10 (2.3) | 33 (4.0) | |
| Black | 191 (43.8) | 181 (22.1) | |
| White | 219 (50.2) | 588 (71.7) | |
| Other (Indian, Middle Eastern, Native American) | 16 (3.8) | 18 (2.3) | |
| Hispanic ethnicity, n (%) | 62 (14.2) | 125 (15.2) | 0.63 |
| Male sex, n (%) | 240 (55.0) | 452 (55.1) | 0.98 |
| Drugs frequently associated with hyponatremia, n (%) | 336 (77.1) | 627 (76.5) | 0.81 |
| Antidepressants | 135 (31.0) | 347 (42.3) | < 0.001** |
| Antiepileptics | 61 (14.0) | 154 (18.8) | 0.03 |
| Thiazides | 16 (3.7) | 54 (6.6) | 0.03 |
| Haloperidol | 27 (6.2) | 168 (20.5) | < 0.001** |
| Proton pump inhibitor | 257 (59.0) | 319 (38.9) | < 0.001** |
| Chronic kidney disease | 205 (47.0) | 240 (29.3) | < 0.001** |
| Liver disease | 40 (9.2) | 130 (15.9) | 0.001** |
| Congestive heart failure | 171 (39.2) | 249 (30.4) | 0.002** |
Mean age was tested with independent t-testing. All other variables were frequency data and analyzed by χ² testing. Statistically significant results are denoted at α = 0.01 and signified with **. Participant race, antidepressant use, haldol use, PPI use, CKD proportion, liver disease proportion, and CHF proportion had statistically significant variations between those with and without enteral feeding tubes. Race and Hispanic ethnicity were recorded as separate variables; therefore, participants identified as Hispanic may also be represented within racial categories
Table 2.
Descriptive statistics between groups separated by incident hyponatremia
| Variable | Hyponatremia (n = 435) | No Hyponatremia (n = 821) | p-value |
|---|---|---|---|
| Age, mean (SD) | 66.9 (14.2) | 66.8 (15.2) | 0.89 |
| Race, n (%) | 0.0004** | ||
| Asian | 14 (3.2) | 29 (3.5) | |
| Black | 160 (36.8) | 212 (25.8) | |
| White | 247 (56.8) | 560 (68.2) | |
| Other (Indian, Middle Eastern, Native American) | 14 (3.2) | 20 (2.4) | |
| Hispanic ethnicity, n (%) | 76 (17.5) | 111 (13.5) | 0.06 |
| Male sex, n (%) | 248 (57.0) | 444 (54.1) | 0.32 |
| Drugs frequently associated with hyponatremia, n (%) | 335 (77.0) | 628 (76.5) | 0.84 |
| Antidepressants | 146 (33.6) | 336 (40.9) | 0.01** |
| Antiepileptics | 67 (15.4) | 148 (18.0) | 0.24 |
| Thiazides | 21 (4.8) | 49 (6.0) | 0.40 |
| Haloperidol | 48 (11.0) | 147 (17.9) | 0.001** |
| Proton pump inhibitor | 232 (53.3) | 344 (41.9) | < 0.001** |
| Chronic kidney disease | 245 (56.3) | 200 (24.4) | < 0.001** |
| Liver disease | 66 (15.2) | 104 (12.7) | 0.22 |
| Congestive heart failure | 194 (44.6) | 226 (27.5) | < 0.001** |
Mean age was tested with independent t-testing. All other variables were frequency data and analyzed by χ² testing. Statistically significant results are denoted at α = 0.01 and signified with **. Participant race, antidepressant use, haldol use, PPI use, CKD proportion, and CHF proportion had statistically significant variations between those with and without incident hyponatremia. Race and Hispanic ethnicity were recorded as separate variables; therefore, participants identified as Hispanic may also be represented within racial categories
Multivariable analysis
A multivariate modified Poisson regression model with robust variance estimation was constructed to evaluate independent associations between clinical predictors and incident hyponatremia. After adjustment for demographic and clinical covariates, enteral feeding tube use was strongly associated with incident hyponatremia (adjusted RR 2.42; 95% CI 1.84–3.19; p < 0.0001). Chronic kidney disease (CKD) was independently associated with increased risk of hyponatremia (adjusted RR 1.87; 95% CI 1.43–2.44; p < 0.0001). Liver disease was also significantly associated with increased risk (adjusted RR 1.58; 95% CI 1.20–2.07; p = 0.001). Both congestive heart failure (adjusted RR 1.30; 95% CI 1.00–1.69; p = 0.052) and Hispanic ethnicity (adjusted RR 1.25; 95% CI 1.04–1.51; p = 0.019) approached statistical significance but did not achieve the predetermined statistical significance (α = 0.01). These results are summarized in Table 3. Age, sex, and each of the high-risk medications were not significantly associated with incident hyponatremia in the adjusted model.
Table 3.
Stratified associations between clinical predictors and incident hyponatremia by enteral feeding tube status
| Predictor | Adjusted RR | 95% CI | p-value |
|---|---|---|---|
| Enteral feeding tube (Yes; ref No) | 2.42 | 1.84–3.19 | < 0.0001** |
| Chronic kidney disease (Yes; ref No) | 1.87 | 1.43–2.44 | < 0.0001** |
| Liver disease (Yes; ref No) | 1.58 | 1.20–2.07 | 0.001** |
| Congestive heart failure (Yes; ref No) | 1.30 | 1.00–1.69 | 0.052 |
| Hispanic ethnicity | 1.25 | 1.04–1.51 | 0.019 |
| Male sex | 1.06 | 0.92–1.22 | 0.403 |
| Age (per year) | 0.999 | 0.994–1.004 | 0.710 |
| Antidepressant use (Yes; ref No) | 0.92 | 0.79–1.07 | 0.291 |
| Antiepileptic use (Yes; ref No) | 1.11 | 0.90–1.36 | 0.322 |
| Thiazide use (Yes; ref No) | 0.95 | 0.68–1.34 | 0.786 |
| Haldol use (Yes; ref No) | 1.02 | 0.80–1.30 | 0.892 |
| PPI use (Yes; ref No) | 0.99 | 0.86–1.15 | 0.923 |
| Interaction Terms | |||
| Enteral feeding tube × CKD | 1.06 | 0.77–1.47 | 0.704 |
| Enteral feeding tube × CHF | 0.85 | 0.62–1.15 | 0.295 |
| Enteral feeding tube × Liver disease | 0.72 | 0.50–1.04 | 0.083 |
Modified Poisson regression with log link and robust variance estimation (GEE). Adjusted relative risks (RRs) are reported. Statistically significant predictors for incident hyponatremia were enteral feeding tubes, CKD, and liver disease. CHF and hispanic ethnicity showed low p-values. However, these failed to reach the cutoff for statistical significance. Age, sex, and each of the high-risk medications did not show statistical significance. Interaction terms represent multiplicative effects beyond the independent main effects of each variable. No statistically significant interaction was observed between enteral feeding tube status and CKD, CHF, or liver disease on incident hyponatremia. Statistically significant results are denoted at α = 0.01 and signified with **
Interaction terms between enteral feeding tube status and CKD (adjusted RR 1.06; 95% CI 0.77–1.47; p = 0.704), CHF (adjusted RR 0.85; 95% CI 0.62–1.15; p = 0.295), and liver disease (adjusted RR 0.72; 95% CI 0.50–1.04; p = 0.083) were not statistically significant, indicating no evidence of multiplicative effect modification in the adjusted model. The absence of significant interaction terms suggests that the association between enteral feeding tube use and cumulative incident hyponatremia was relatively consistent across comorbidity strata.
Discussion
Principal findings
In this retrospective cohort of SNF residents, incident hyponatremia was common, affecting approximately one-third of participants. The mean age of the cohort was somewhat younger than in many geriatric SNF studies; however, the population studied reflects the increasingly diverse case mix of modern skilled nursing facilities, which frequently include younger adults with complex medical needs. Enteral feeding tube use was strongly associated with cumulative incident hyponatremia. Residents with enteral feeding tubes demonstrated more than double the observed cumulative risk compared with those without enteral feeding tubes. CKD and liver disease were also independently associated with increased risk. CHF, age, sex, ethnicity, and high-risk medication exposure were not independently associated with incident hyponatremia after adjustment. The cumulative incidence of hyponatremia observed in this cohort is consistent with these prior observations [1–7] and reinforces the clinical importance of electrolyte monitoring in long-term care settings.
The absence of a medication effect may reflect exposure misclassification, residual confounding, or the high baseline medical complexity of this population. The absence of significance from age, sex, and ethnicity suggests these demographic variables did not independently contribute to incident hyponatremia after adjustment. CHF did not demonstrate a statistically significant association with incident hyponatremia at the predetermined conservative alpha threshold, although a type II error remains possible. CKD and liver disease remained independently associated with incident hyponatremia after adjustment, findings that are biologically plausible given their known effects on fluid regulation and renal water handling. No significant interaction effects were observed between enteral feeding tube status and major comorbid conditions.
Enteral feeding tube use, CKD, and liver disease were independently associated with incident hyponatremia, with enteral feeding tube use and CKD showing larger adjusted effect estimates in this cohort. In CKD, impaired renal free water clearance likely contributes to sodium dysregulation [1–2, 6, 15–16, 19–20]. In liver disease, alterations in effective circulating volume and fluid regulation may promote dilutional hyponatremia [1–2, 19–20]. CHF may similarly contribute through neurohormonal activation and reduced effective arterial perfusion, although this association did not reach statistical significance in the present study. The lack of statistical significance for the interaction terms suggests enteral feeding tube status did not substantially modify the relationship between these comorbid conditions and incident hyponatremia.
Several physiologic mechanisms may plausibly contribute to the observed association between enteral feeding tube use and incident hyponatremia. Residents receiving enteral nutrition frequently experience reduced dietary solute intake relative to oral feeding patterns, particularly when enteral formulas contain relatively low sodium content or when additional free water flushes are administered to maintain hydration and tube patency. In the setting of low solute intake, renal free water excretion capacity may become impaired, particularly among medically complex older adults with reduced renal reserve or concurrent comorbid disease. Reduced obligate solute intake may limit maximal renal free water excretion capacity, thereby increasing susceptibility to dilutional hyponatremia. Additionally, enteral feeding tube recipients often exhibit underlying neurologic impairment, frailty, chronic illness, and malnutrition, all of which may further contribute to impaired water balance regulation and increased vulnerability to electrolyte abnormalities. Variability in enteral feeding regimens, sodium content, and total fluid administration may therefore represent important contributors to the observed association.
Comparison with prior literature
The findings of this study are consistent with prior literature outlining hyponatremia as a common electrolyte disorder among older adults in SNFs. Previous studies have estimated hyponatremia prevalence in skilled nursing facilities between 15% and 30%, typically higher than estimates reported among community-dwelling older adults [1–7]. Several chronic conditions previously associated with hyponatremia were also identified in this analysis. The observed associations with CKD and liver disease are consistent with prior epidemiologic and physiologic literature describing impaired water excretion, altered renal sodium handling, and neurohormonal dysregulation in these conditions [1–2, 6–10, 19]. These findings align with previous studies demonstrating increased hyponatremia risk among patients with renal dysfunction, advanced liver disease, and other conditions associated with volume dysregulation.
In contrast to themes in the literature, CHF and high-risk medication exposure were not independently associated with incident hyponatremia. Although these factors are frequently cited as contributors to hyponatremia in hospitalized populations, their effects in long-term care residents may be more heterogeneous and influenced by underlying disease severity, medication dosing, or clinical monitoring patterns. It is also possible that the high baseline medical complexity of the cohort reduced the relative contribution of individual medication classes or conditions.
Relatively few studies have evaluated the relationship between enteral feeding and hyponatremia in long-term care settings. Case reports and smaller clinical studies have suggested that enteral nutrition through enteral tube feeding may influence electrolyte balance through mechanisms such as free-water administration, dilutional effects of feeding regimens, or underlying comorbid illness [1, 5, 16]. The strong association observed between enteral feeding tube use and incident hyponatremia in this cohort extends these observations and suggests that residents receiving enteral nutrition may represent a subgroup at elevated risk.
These findings support prior evidence linking chronic disease burden to hyponatremia while highlighting enteral feeding tube use as a potentially important and understudied risk factor in skilled nursing facility residents. These findings also suggest that care-related exposures, including enteral feeding practices, may warrant consideration when evaluating electrolyte abnormalities in institutionalized populations.
Clinical implications
Although enteral feeding tube use in SNF settings is associated with substantially higher risk of incident hyponatremia, comorbid conditions such as CKD, and liver disease further increase this risk. Residents with enteral feeding tubes, CKD, or liver disease demonstrated increased cumulative risk of incident hyponatremia. These findings support consideration of closer electrolyte monitoring among residents with enteral feeding tubes, CKD, or liver disease.
Free water flush prescribing practices for enteral feeding tubes are often individualized and may vary substantially between physicians, nurse practitioners, dietitians, and nursing or institutional protocols. In the absence of standardized enteral hydration protocols, differences in prescribed flush volumes and frequencies may contribute to variability in sodium balance and risk of dilutional hyponatremia among SNF residents, which may contribute to variability in sodium balance among residents with enteral feeding tubes.
Additionally, actual free water exposure may differ from prescribed enteral hydration orders because bedside nursing practices during medication administration, tube patency maintenance, and additional discretionary flushing were not standardized or measurable within the retrospective dataset. Consequently, cumulative free water administration may have exceeded documented ordered volumes in some residents. Standardization and adherence to prescribed enteral free water flushing protocols could not be evaluated and may represent an important unmeasured source of variability in free water exposure.
Clinicians and public health professionals caring for SNF residents should consider developing or refining risk management strategies to identify individuals at highest risk. In the absence of a validated hyponatremia-specific screening tool, existing instruments such as the Identification of Seniors at Risk tool may be considered. Additionally, emerging risk stratification approaches from the STRATIFY study may offer a hyponatremia-specific screening framework once publicly available. Implementation of such tools could potentially reduce incident hyponatremia, associated adverse events, hospitalizations, and mortality [4, 6–14, 16, 18–20, 22–23, 25]. As aging populations increasingly rely on institutional long-term care globally, identifying modifiable risk factors for electrolyte disturbances carries broader clinical and public health implications.
Strengths
This study has several strengths. The cohort included more than 1,200 skilled nursing facility residents with available laboratory data, providing a relatively large sample for examining electrolyte abnormalities in a long-term care setting. Incident hyponatremia was identified using objective serum sodium measurements obtained during routine clinical care rather than administrative coding, which reduces the potential for outcome misclassification. Importantly, the study evaluates enteral feeding tube use as a potential risk factor for hyponatremia, an exposure that has received limited attention in studies of skilled nursing facility residents.
Because laboratory testing reflected routine clinical practice rather than protocol-driven monitoring, the findings may better represent real-world detection patterns in skilled nursing facilities. The study also evaluated multiple clinically relevant comorbid conditions and medication exposures within the same analytic framework, allowing comparison of several commonly cited risk factors in this population. Finally, the use of modified Poisson regression allowed direct estimation of relative risks for a common outcome, providing effect estimates that are more interpretable for clinical and epidemiologic research in skilled nursing facility populations.
Limitations
Several limitations should be considered when interpreting these findings. This was a single-site study, which may limit generalizability. The study facility receives specialized referrals including ventilator-dependent patients and in-house dialysis, which may reflect a population with higher medical complexity than typical community SNFs. The inclusion of dialysis-dependent residents may further limit generalizability, as these patients may differ substantially from typical skilled nursing facility populations and may possess distinct baseline risks for electrolyte abnormalities including hyponatremia. The retrospective cohort design introduces potential for selection bias, misclassification bias, surveillance bias, and unmeasured confounding. Residents receiving enteral feeding often require more intensive clinical monitoring and may undergo more frequent laboratory testing, which could increase the likelihood of detecting hyponatremia relative to residents without feeding tubes. Because laboratory monitoring was clinician-directed rather than standardized, residents with greater medical complexity may have undergone more frequent sodium testing. Laboratory testing frequency was determined by treating clinicians according to resident acuity and clinical judgment.
Detailed information regarding enteral free water flushes, feeding regimens, and concurrent oral intake was not consistently available and therefore could not be evaluated. Additionally, detailed information regarding enteral feeding formula composition, sodium concentration, and total administered fluid volume was not consistently available. Consequently, the observed association between enteral feeding tube use and incident hyponatremia may partially reflect differences in solute intake, sodium content, or fluid administration practices rather than enteral feeding tube status alone. Nutritional status and malnutrition severity were also not directly assessed. Reliable nutritional assessment data were inconsistently available, and body mass index may not accurately reflect nutritional status in medically complex skilled nursing facility populations due to edema, fluid shifts, immobility, and chronic illness. Future studies incorporating standardized nutritional assessments may help clarify the relationship between malnutrition, enteral feeding, and hyponatremia risk.
Reliance on existing electronic health record data may have increased the likelihood of exposure or outcome misclassification, potentially biasing estimates toward the null. Follow-up duration and laboratory monitoring frequency were not standardized and may have varied according to clinical complexity. Residents with enteral feeding tubes may have experienced longer facility stays and more frequent serum sodium testing, increasing the likelihood of detecting hyponatremia. Because time-to-event modeling was not feasible, the outcome was analyzed as cumulative incidence during the available observation period. Because residents receiving enteral feeding frequently undergo more intensive laboratory monitoring and prolonged institutional care, differential surveillance may have substantially contributed to the magnitude of the observed association.
Residual confounding remains possible, and medication adherence could not be assessed due to the data that was available. Medication exposures were treated as cumulative, which can obscure true causal associations and these variables were mainly used to determine any true baseline differences between main exposure and outcomes. Combining heterogeneous medication classes into a single exposure variable may have obscured medication-specific associations. The definition of incident hyponatremia included mild cases that may be clinically insignificant; future studies may consider a lower sodium threshold (e.g., < 130 mEq/L) or focus on symptomatic and clinically significant hyponatremia cases. The duration, severity, recurrence, and treatment of hyponatremia episodes were not consistently available for standardized analysis and therefore were not incorporated into the study design. Some subgroup analyses were limited by small sample sizes, reducing precision. Demographic subgroup analyses were not performed, as they were beyond the scope of the study objectives. Duration of follow-up and frequency of sodium testing were not available for quantification and therefore could not be incorporated into the analytic model. As a result, differences in length of stay or monitoring intensity between exposure groups may have influenced the observed cumulative risk estimates. These findings should therefore be interpreted as associations within an available observation window rather than precise time-dependent risk estimates.
Conclusion
Hyponatremia remains a prevalent issue in older adults in SNF settings. Prior literature suggests that recognition and management of hyponatremia may improve motor and cognitive function, reducing morbidity and enhancing quality of life [10, 11]. Residents with enteral feeding tubes are at greater risk than those without enteral feeding tubes. CKD and liver disease independently contributed to increased risk, while CHF demonstrated a nonsignificant trend toward increased risk. High-risk medications did not demonstrate an independent association in this cohort. No significant multiplicative interactions were observed between enteral feeding tube status and major chronic comorbidities. Findings of this study support the need for healthcare and public health practitioners to consider enhanced monitoring protocols for high-risk patients, particularly those with enteral feeding tubes and significant chronic comorbidities. Understanding and recognizing high-risk residents may help support earlier identification and management of hyponatremia.
Acknowledgements
The authors declare no conflicts of interest and no source of funding for this project. The authors used ChatGPT (OpenAI) to assist with language editing and readability. The tool was not used for data analysis, study design, or interpretation of results. All analytic decisions and conclusions are those of the authors. After using this tool/service, the author(s) reviewed and edited the content as needed and take(s) full responsibility for the content of the published article. The authors would like to thank the IRB of the hosting organization as well as the skilled nursing facility that hosted data collection.
Author contributions
Conceptualization: D.M., T.M., and I.R.Data curation: D.M.Methodology: D.M., T.M., and I.R.Formal analysis: D.M. and E.H.Investigation: D.M., T.M., and I.R.Writing—original draft: D.M. and E.H.Writing—review and editing: D.M., T.M., I.R., and E.H.Supervision: D.M.All authors read and approved the final manuscript.
Funding
This research received no external funding.
Data availability
The datasets generated and/or analyzed during the current study are not publicly available due to data use agreements and privacy restrictions but may be available from the corresponding author on reasonable request and with permission of the data-holding institution.
Declarations
Ethics approval and consent to participate
Ethics approval for this study was obtained from the Rosalind Franklin University of Medicine and Science Institutional Review Board. The requirement for informed consent was waived due to the retrospective design and use of de-identified data. All methods were performed in accordance with relevant guidelines and regulations, including the ethical principles outlined in the Declaration of Helsinki.
Consent for publication
Not applicable.
Competing interests
The authors declare no competing interests.
Footnotes
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
References
- 1.Filippatos TD, Makri A, Elisaf MS, Liamis G. Hyponatremia in the elderly: challenges and solutions. Clin Interv Aging. 2017;12:1957–65. 10.2147/CIA.S138535. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Kugler JP, Hustead T. Hyponatremia and hypernatremia in the elderly. Am Fam Physician. 2000;61:3623–30. [PubMed] [Google Scholar]
- 3.Mannesse CK, Vondeling AM, van Marum RJ, van Solinge WW, Egberts TCG, Jansen PAFJ. Prevalence of hyponatremia on geriatric wards compared to other settings over four decades: a systematic review. Ageing Res Rev. 2013;12:165–73. 10.1016/j.arr.2012.04.006. [DOI] [PubMed] [Google Scholar]
- 4.Miller M, Morley JE, Rubenstein LZ. Hyponatremia in a nursing home population. J Am Geriatr Soc. 1995;43:1410–3. 10.1111/j.1532-5415.1995.tb06623.x. [DOI] [PubMed] [Google Scholar]
- 5.Serrano Valles C, Gómez Hoyos E, Ortolá Buigues A, Primo Martín D, Jiménez Sahagún R, García Calvo S, et al. Prevalencia y factores asociados a la hiponatremia en pacientes con nutrición enteral [Hyponatremia among patients with total enteral tube feeding: prevalence and associated clinical factors]. Nutr Hosp. 2022;39:723–7. 10.20960/nh.03964. [DOI] [PubMed] [Google Scholar]
- 6.Fratangelo L, Nguyen S, D’Amelio P. Hyponatremia and aging-related diseases: key player or innocent bystander? A systematic review. Syst Rev. 2023;12:84. 10.1186/s13643-023-02246-w. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Aydin AE, Ates Bulut E, Kocyigit SE, Dost FS, Mutlay F, Altunkalem Seydi K, et al. An instant relationship between hyponatremia, geriatric syndromes, and drugs in older adults: a cross-sectional analysis from a single geriatric clinic. Diagnostics (Basel). 2025;15:744. 10.3390/diagnostics15060744. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Ande SP, Kanitkar S, Borle A, Ahlawat M. Hyponatremia among elderly hospitalized patients: an observational study. Cureus. 2024;16:e67632. 10.7759/cureus.67632. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Shi LT, Feng Z, Zhu CM. A retrospective study: exploring preoperative hyponatremia in elderly patients with hip fractures. J Orthop Surg Res. 2024;19:186. 10.1186/s13018-024-04643-w. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Kapoor M, Dhar M, Pathania M. The assessment of baseline comprehensive geriatric assessment parameters in geriatric patients with varying severity of hyponatremia at a tertiary care center. Cureus. 2022;14:e21516. 10.7759/cureus.21516. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Kapoor M, Pathania M, Dhar M. Serum sodium improvement: change in comprehensive geriatric assessment parameters in geriatric patients with hyponatremia. BMC Geriatr. 2023;23:666. 10.1186/s12877-023-04299-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Netzer S, Gastens V, Boland B, et al. Association between hyponatremia and mortality and readmission in multimorbid older adults: a cohort study. J Clin Med. 2023;14:7146. 10.3390/jcm14207146. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Boyer S, Gayot C, Bimou C, et al. Prevalence of mild hyponatremia and its association with falls in older adults admitted to an emergency geriatric medicine unit (the MUPA Unit). BMC Geriatr. 2019;19:265. 10.1186/s12877-019-1282-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Ioannou P, Panagiotakis S, Tsagkaraki E, Tsioutis C, Fragkiadakis K, Gikas A, et al. Increased mortality in elderly patients admitted with hyponatremia: a prospective cohort study. J Clin Med. 2021;10:3059. 10.3390/jcm10143059. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Su Y, Ma M, Zhang H, Pan X, Zhang X, Zhang F, et al. Prognostic value of serum hyponatremia for outcomes in patients with heart failure with preserved ejection fraction: an observational cohort study. Exp Ther Med. 2020;20:101. 10.3892/etm.2020.9231. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Djukic M, Grewe J, Kunz O, Gross O, Nau R. Hyponatremia in geriatric patients. Z Gerontol Geriatr. 2024. 10.1007/s00391-024-02342-z. [DOI] [PubMed] [Google Scholar]
- 17.Goto M, Sakai M. Proton-pump inhibitor use and hyponatremia. Nihon Ronen Igakkai Zasshi. 2023;60:153–7. 10.3143/geriatrics.60.153. [DOI] [PubMed] [Google Scholar]
- 18.Jun K, Kim Y, Ah YM, Lee JY. Awareness of the use of hyponatraemia-inducing medications in older adults with hyponatraemia: a study of their prevalent use and association with recurrent symptomatic or severe hyponatraemia. Age Ageing. 2021;50:1137–43. 10.1093/ageing/afaa195. [DOI] [PubMed] [Google Scholar]
- 19.Schrier RW, Abraham WT, Koomans HA, et al. Hyponatremia in the elderly: epidemiology, diagnosis, and management in nursing home patients. J Am Geriatr Soc. 1995;43:547–52. 10.1111/j.1532-5415.1995.tb07891.x. [DOI] [Google Scholar]
- 20.Baser S, Yılmaz CN, Gemcioglu E. Do the etiology of hyponatremia and serum sodium levels affect the length of hospital stay in geriatric patients with hyponatremia? J Med Biochem. 2022;41:40–6. 10.5937/jomb0-29914. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Simakoloyi N, Erasmus E, van Hoving DJ. The characteristics of geriatric patients managed within the resuscitation unit of a district-level emergency centre in Cape Town. Afr J Emerg Med. 2022;12:39–43. 10.1016/j.afjem.2021.11.005. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Thorpe O, Cuesta M, Fitzgerald C, Feely O, Tormey WP, Sherlock M, et al. Active management of hyponatraemia and mortality in older hospitalised patients compared with younger patients: results of a prospective cohort study. Age Ageing. 2021;50:1144–50. 10.1093/ageing/afaa248. [DOI] [PubMed] [Google Scholar]
- 23.Rajarajeswari TB, Arul Senghor KA, Vinodhini VM, Kumar JS, Prasath N. Hyponatremia and Identification of Seniors at Risk (ISAR) score in geriatric patients: an analytical cross-sectional study. Cureus. 2023;15:e49493. 10.7759/cureus.49493. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Hillier TA, Abbott RD, Barrett EJ. Hyponatremia: evaluating the correction factor for hyperglycemia. Am J Med. 1999;106:399–403. 10.1016/S0002-9343(99)00055-8. [DOI] [PubMed] [Google Scholar]
- 25.Wang A, Koshiaris C, Archer L, et al. Developing prediction models for electrolyte abnormalities in patients indicated for antihypertensive therapy: evidence-based treatment and monitoring recommendations. J Hypertens. 2025;43:1348–59. 10.1097/HJH.0000000000004032. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
The datasets generated and/or analyzed during the current study are not publicly available due to data use agreements and privacy restrictions but may be available from the corresponding author on reasonable request and with permission of the data-holding institution.