ELECTROHEALTH study: prevalence and prognostic significance of mild hypophosphatemia in older adults

Electrolyte imbalances are common in older hospitalized patients and are increasingly recognized as important determinants of clinical outcomes. These alterations, often subtle and underdiagnosed, may contribute to functional decline, prolonged hospitalization, early rehospitalization, and increased mortality. While hyponatremia and hypokalemia have been extensively investigated [1‐4] hypophosphatemia remains relatively understudied in the geriatric population, despite its potential clinical implications.

Phosphorus is an essential mineral involved in several biochemical and physiological processes, including nucleic acid synthesis, enzymatic activity, cell membrane integrity, neurological function, and skeletal mineralization [5]. Approximately 85% of body phosphorus is stored in the skeleton, while the remainder is distributed in soft tissues and extracellular fluids. Serum phosphate concentrations, although representing only a small fraction of total body stores, are widely used as a clinical marker of phosphorus status. Normal serum phosphate levels range from 2.5 to 4.5 mg/dL (0.81 to 1.45 mmol/L); hypophosphatemia is generally defined as serum phosphate below 2.5 mg/dL (0.81 mmol/L), with severe hypophosphatemia defined at levels below 1 mg/dL (0.32 mmol/L).

Severe hypophosphatemia is well known to cause acute and potentially life-threatening complications such as respiratory muscle weakness, osteomalacia, impaired platelet function, hemolysis, seizures, and heart failure. By contrast, non-severe hypophosphatemia is often considered clinically silent and therefore overlooked [6, 7]. Nevertheless, mild phosphate deficiency in older patients has been associated to negative health outcomes; in a recent systematic review from our group, we suggested that mild hypophosphatemia (HP) may be considered not only as a marker of poor prognosis but also as a possible contributing factor to clinical deterioration [8].

Disturbances in other minerals may coexist and interact with phosphate metabolism. Hypocalcemia, for instance, has been associated with increased cardiovascular and all-cause mortality in community and hospitalized populations, suggesting that broader alterations in mineral homeostasis may carry prognostic implications [9, 10].

Recent studies also point to hypomagnesemia as a potential predictor of adverse outcomes in geriatric patients [11] raising the possibility that magnesium deficiency could confound the association between phosphate levels and clinical outcomes.

Despite the associations between mild HP and negative clinical outcomes as frailty, cognitive impairment, depressed mood, increased mortality and higher risk of falls during hospitalization [12‐14], significant knowledge gaps remain. The available studies are heterogeneous in design, sample size, and diagnostic thresholds for hypophosphatemia. Furthermore, most studies did not adjust for relevant confounding factors such as multimorbidity, polypharmacy, and nutritional status, all of which are highly prevalent in older adults and may influence phosphate levels. Therefore, it is still unclear whether mild HP represents an independent risk factor for poor outcomes or merely reflects underlying comorbidity and frailty.

Addressing this uncertainty is particularly relevant in geriatric inpatients, who often present multiple risk factors for electrolyte imbalance, including inadequate nutritional intake, polypharmacy, renal impairment, and acute illness. Within this context, the ELECTROHEALTH study was designed to investigate the role of electrolyte imbalances—including hyponatremia, hypokalemia, and hypophosphatemia—in predicting adverse health outcomes among older inpatients [15].

In the present analysis, we specifically focus on the prognostic significance of serum phosphate levels measured during hospital stay. The primary objective was to examine whether serum phosphate levels are independently associated with one‑year mortality.

The secondary objectives were to evaluate the associations between phosphate levels and the following outcomes: length of hospital stay, functional decline, cognitive decline, 30‑day rehospitalization, and one‑year rehospitalization.

Our cohort comprised 1,287 patients of those 32 have hyperphosphatemia and 1 severe hypophosphatemia, those patients were not analyzed leaving a cohort of 1,254 patients for the analyses, of whom 190 (15.1%) presented with mild HP and 1,064 (84.9%) had normal phosphate levels. The two groups were comparable in age and showed no significant differences in sex distribution (female: 58.9% vs. 65.9%, p = 0.065) or mortality (alive at last follow-up: 56.3% vs. 60.7%, p = 0.254). However, rehospitalization was more frequent among patients with mild HP (61.6% vs. 53.5%, p = 0.04), moreover HP patients have lower MMSE score. No significant differences were observed for functional status (Tinetti, IADL, ADL), nutritional risk (NRS-2002), BMI, length of stay, or survival days (Table 1). Among the patients in our cohort, cognitive screening was initiated with the Mini-Cog. A total of 207 patients completed this initial assessment (176 with a negative Mini-Cog and 31 with a positive result). Among those who underwent further cognitive evaluation, 181 patients completed the MMSE, while 83 patients completed the MoCA, according to clinical judgment and the suitability of each tool for the individual patient.

Several biological mechanisms help explain why hypophosphatemia may contribute to adverse clinical outcomes in older adults. Phosphate is essential for ATP generation, neuromuscular function, membrane integrity, and oxygen delivery through its impact on 2,3-Diphosphoglicerate levels in red blood cells [24]. Low phosphate availability has been linked to impaired muscle performance and decreased ATP synthesis [25], which may promote weakness and increase fall risk, as demonstrated by studies showing more than a two-fold higher fall rate among patients with low serum phosphate. In addition, hypophosphatemia may compromise immune function by reducing phagocytosis and chemotaxis, thereby predisposing individuals to infections such as pneumonia and contributing to more severe disease courses and prolonged hospital stays [8]. All these considerations make even non-severe reductions possibly clinically relevant in physiologically vulnerable patients.

Building on these biological mechanisms, our epidemiological study is consistent with a population that is particularly susceptible to transient phosphate derangements. In our cohort, the prevalence of mild HP was 15.1%, slightly higher than the 12.5% reported in our systematic review of 26,548 patients [8]. This difference may reflect the characteristics of our hospitalized population, including acute illness and multimorbidity—both known contributors to electrolyte imbalances. The weighted comorbidity burden was moderate, indicating substantial chronic disease complexity, which may predispose patients to HP through mechanisms such as polypharmacy and organ dysfunction. Variability in HP definitions across studies (cut-offs ranging from 2.0 to 3.0 mg/dL) may also account for differences in prevalence estimates. Furthermore, the review highlighted a strong association between HP and frailty, suggesting that our cohort represents a particularly vulnerable subgroup. Previous studies have linked HP to diverse clinical outcomes, including electrolyte imbalance [12], influenza in older patients [26], falls during hospitalization [13], pneumonia prognosis [23, 27], outcomes after coronary artery bypass grafting [14], and malnutrition [28], underscoring its multifaceted clinical implications.

Regarding adverse outcomes, our findings partially align with the literature. Mild HP was significantly associated with rehospitalization, whereas mortality, functional status, nutritional risk, and length of stay did not differ between groups. MMSE scores were lower in the HP group, suggesting a possible link between phosphate levels and cognitive function, consistent with prior evidence associating HP with cognitive impairment, depressed mood, and fall risk. However, in adjusted models, phosphate was not an independent predictor of falls or cognitive decline, indicating that age remains the dominant factor.

Discharge destinations are also associated with HP: patients transferred to higher-level acute units or nursing homes had a higher prevalence compared to those discharged home or to rehabilitation. This association disappeared when phosphate was analyzed as a continuous variable, suggesting it reflects the binary HP definition rather than a true shift in phosphate distribution. These findings likely mirror greater frailty and multimorbidity among patients requiring institutional care.

Unlike Pourhassan et al. [28], who reported a strong association between HP and malnutrition (86% vs. 56% in non-HP patients), we found no significant differences in nutritional risk (NRS-2002) or BMI. This discrepancy may relate to differences in assessment tools, population characteristics, or HP severity thresholds. Our results suggest that mild HP alone may not be a reliable marker of malnutrition risk in this setting.

A major strength of our study is the large, well-characterized cohort and the use of comprehensive geriatric assessments. However, limitations include its retrospective design that limits causal inference and relies on routinely collected data. The exact timing of phosphate measurements relative to admission was not available, introducing potential variability. Important confounders such as serum magnesium, calcium, vitamin D status, and viral infections could not be included due to incomplete data. Furthermore, the single‑center setting may limit generalizability. As regards cognitive assessment, patients with a positive Mini‑Cog were then evaluated in depth with either the MMSE or MoCA, depending on clinical judgment and the patient’s functional or linguistic profile. Because only Mini‑Cog–positive patients proceeded to second‑level testing, individuals with normal Mini‑Cog results were not systematically assessed with MMSE or MoCA. This stepped diagnostic pathway may introduce selection bias in the distribution of cognitive scores; however, it reflects the standard clinical approach in which detailed cognitive testing is reserved for patients who do not pass the initial screening. Although mild HP was not independently predictive of major outcomes, its association with frailty markers and rehospitalization suggests it may serve as a clinical indicator of vulnerability rather than a direct causal factor. This distinction has implications for clinical monitoring but does not support routine supplementation in asymptomatic mild HP.

In fact, the clinical management of non-severe HP remains controversial. Current guidelines recommend addressing underlying causes as the primary strategy, with supplementation generally unnecessary for mild HP unless symptoms or ongoing losses are present. When indicated, oral phosphate is preferred for moderate reductions, while intravenous replacement is reserved for severe or symptomatic cases. Given the potential adverse effects of supplementation and the lack of evidence for benefit in non-severe HP, a conservative, individualized approach appears most appropriate. This is particularly relevant in our cohort, characterized by frailty, multimorbidity, and a higher prevalence of HP among patients discharged to institutional care, where the risks of overtreatment may outweigh uncertain benefits.

In summary, non-severe HP is a relatively common yet underrecognized condition in older hospitalized adults. We observed associations with frailty, cognitive impairment, and rehospitalization, but whether HP is a causal factor, or a marker of vulnerability remains unclear. Further large-scale, prospective studies are needed to clarify its prognostic value and determine whether targeted interventions can improve outcomes in this population.