Age and Ageing Advance Access originally published online on November 4, 2008
Age and Ageing 2009 38(1):112-115; doi:10.1093/ageing/afn244
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Associations and consequences of hypophosphataemia in older hospitalised women
SIR—Hypophosphataemia has been associated with increased morbidity and mortality in a number of patient groups, including patients on intensive care units [1], patients with respiratory tract infections [2], severe sepsis [3] and older patients [4]. Low phosphate has a number of adverse physiological effects, including myocardial depression, neuromuscular weakness, rhabdomyolysis, haemolysis, impairment of white blood cell function and decreased insulin sensitivity.Most studies examining the association between low phosphate and mortality have focussed on patients with severe hypophosphataemia, sometimes defined as serum phosphate of <0.32 mmol/l (1 mg/dl) [5]. Very few studies have examined the effect of more modest degrees of hypophosphataemia or the occurrence of hypophosphataemia in older hospitalised patients [4]. Impaired homeostasis and lack of physiological reserve in old age suggest that older patients may not only be at high risk of hypophosphataemia but could be more susceptible to its adverse effects.
The aim of this study was to examine the prevalence of hypophosphataemia in a cohort of hospitalised older women, determine factors associated with low phosphate and examine the effect of hypophosphataemia on outcome in terms of mortality and length of hospital stay.
We studied consecutive admissions to a single Medicine for the Elderly assessment ward at Ninewells Hospital, Dundee, Scotland, from November 2004 to April 2005 inclusive. Female patients are admitted via the emergency department and medical admissions unit and undergo comprehensive geriatric assessment. From November 2004, ward policy was changed to perform magnesium and phosphate measurements on all new admissions to the ward. Tayside local ethics committee deemed that the study did not require formal ethics approval as it utilised existing data collected in routine clinical practice. Approval from the local statutory Data Protection Officer was obtained to access and use confidential patient data.
We prospectively collected biochemistry results on all admissions to the ward, and then performed retrospective notes review on the cohort a mean of 16 (range 7–28) months later. Data were abstracted from the notes by the authors using a standard proforma. Length of stay and date of death were also abstracted from the casenotes; death date is recorded on casenotes even if death occurs in the community. Charlson comorbidity score was calculated from casenote diagnoses [6]. Cognitive impairment was defined as an abbreviated mental test score of <8/10 [7] or a Folstein mini-mental test score of <25/30 [8].
Potassium and magnesium are important intracellular ions and low serum levels are commonly found in older people [9]. Disorders such as refeeding syndrome may cause serum levels of all three ions to fall. Potassium ion activity was measured indirectly by ion selective electrode on the core unit of the Roche P800 modular system. Magnesium was measured using a xylidyl blue colourimetric end point methodology and phosphate was measured by ammonium molybdate to phosphomolybdate as an end point method (all Roche Diagnostics, Lewes, East Sussex, UK). The coefficient of variation for the phosphate assay was 2.3%.
Data were analysed using SPSS version 15 (SPSS, Chicago, USA). Hypokalaemia was defined as a potassium level of <3.5 mmol/l. Hypomagnesaemia was defined as a magnesium level of <0.7 mmol/l. Hypophosphataemia was defined as a phosphate level of <0.8 mmol/l. These cutoffs are those used by our local laboratory as population normal ranges. Simple descriptive statistics were derived for the frequency of low electrolyte levels. Kaplan–Meier curves were constructed to examine the influence of low versus normal electrolyte levels on survival, with the log-rank test used to compare strata. Cox regression analysis, using forward conditional entry, was used to conduct multivariate analysis of factors associated with survival. Because of the small numbers of patients in the study, entry into the multivariate model was conditional on a P value of <0.2. Length of stay was not normally distributed; the Mann–Whitney U test was therefore used to compare length of stay data. Baseline factors associated with low phosphate levels were identified (P < 0.2) and entered into a multivariate binary logistic regression model (P = 0.2 to enter; forward conditional).
One hundred thirty-four patients were admitted during the study period, of whom 125/134 (93%) had retrievable clinical data. Details of these 125 patients are given in Table 1. 96/125 (77%) of patients had phosphate levels checked at least once during their admission. Similarly, 96/125 (77%) of patients had their magnesium levels checked at least once. 124/125 (99%) of patients had potassium levels checked on at least one occasion. 27/96 (29%) of patients had at least one instance of hypophosphataemia (serum phosphate <0.8 mmol/l) during admission. Of these, two had phosphate levels <0.5 mmol/l (moderate hypophosphataemia) and none had levels <0.32 mmol/l (severe hypophosphataemia). 28/96 (29%) had hypomagnesaemia and 40/124 (32%) had hypokalaemia.
|
Phosphate and outcome
18/69 (26.0%) of patients with normal phosphate died during the follow-up compared to 16/27 (59.2%) of those with at least one low phosphate measurement (P = 0.002; Pearsons test). Thirty-day mortality was 5/69 (7.2%) for patients with normal phosphate versus 5/27 (18.5%) for those with at least one low phosphate measurement (P = 0.11; Fisher exact test). Figure 1 shows the difference in mortality between patients with at least one low phosphate measurement compared with those who did not record a low phosphate at any stage. Low magnesium (<0.7 mmol/l) and low potassium (<3.5 mmol/l) were not associated with a significant difference in mortality (P = 0.52 and P = 0.98, respectively, log rank test). The median length of stay was not different in patients with low phosphate and those without a low reading (33 vs. 29 days, P = 0.59; Mann–Whitney U test). We conducted Cox regression analysis to ascertain which baseline factors were independently associated with mortality. Low phosphate, use of intravenous fluids, age, creatinine, Charlson score, peak C reactive protein, active cancer, diabetes and vomiting were all associated with mortality on univariate analysis (P < 0.2). Notably, BMI and albumin levels were not associated with increased mortality. On multivariate Cox regression analysis (forward conditional, P<0.2 to enter), vomiting, creatinine, low phosphate, peak C reactive protein and diabetes were independently associated with mortality.
|
Factors associated with hypophosphataemia
Baseline factors that were found to be univariate associates of hypophosphataemia were hypokalaemia (P < 0.01), receipt of intravenous fluids (P < 0.01), taking calcium and vitamin D (P = 0.06), raised CRP (P = 0.07), low body mass index (P = 0.07), active malignancy (P = 0.10), laxative use (P = 0.12), weight loss (P = 0.13) and increasing age (P = 0.17). GFR, vomiting, diarrhoea, Charlson score and diabetes were not associated with low phosphate. Incorporating the factors into a multivariate regression analysis, hypokalaemia, active cancer, low BMI and use of laxatives were found to be independently associated with hypophosphataemia (binary logistic forward conditional model, P < 0.2 to enter). If hypokalaemia was excluded from the model, active cancer, low BMI, receipt of IV fluids and use of laxatives were found to be independently associated with hypophosphataemia.
Our results confirm that mild degrees of hypophosphataemia are common in older hospitalised women. The incidence of hypophosphataemia was twice the 14.1% rate reported from a recent large sample of older hospitalised patients [4]. Severe hypophosphataemia (<0.32 mmol/l) was not encountered in our sample, which is consistent with previously reported rates in general hospital patients of 0.24% [10]. It has been suggested that mild degrees of hypophosphataemia are not associated with significant clinical consequences [5]. In our patient group however, even mild degrees of hypophosphataemia were associated with markedly increased mortality rates. The lack of difference in length of stay may be due to other factors, including funding delays for care home placement, that affect this variable in our locality.
The overlap between factors associated with hypophosphataemia, and factors known to increase the risk of refeeding syndrome [11] noted in our study and previously [4] may partly explain our results. Many older people are undernourished on admission to hospital [12]; consequent low total body phosphate stores may increase their likelihood of developing hypophosphataemia during illness.
Does low phosphate cause increased mortality, or is it merely a marker for severe illness or frailty? Supplementation with phosphate-containing compounds has been shown to improve respiratory muscle function [13, 14] and myocardial function [1, 15], suggesting that hypophosphataemia itself may contribute to increased mortality via derangement of cardiorespiratory function. Interestingly, a recent large observational study [16] has suggested that higher serum phosphate levels within the normal range are associated with an increased risk of cardiovascular events in middle-aged people. This does not necessarily contradict our results; other pathophysiological processes that depress serum phosphate levels such as refeeding may be obscuring this effect in our patient group.
Our study has several weaknesses. It was a retrospective study carried out at a single centre, involved no males and few patients drank large quantities of alcohol. However, the population studied was similar in age and comorbidity to that seen in many UK units caring for older people. Further work is needed to examine the physiological consequences of low phosphate in older patients and to examine whether repletion of mild hypophosphataemia can improve clinical outcomes.
- Mild degrees of hypophosphataemia are common in older hospitalised women.
- Risk factors for mild hypophosphataemia overlap with those known to precipitate the refeeding syndrome.
- Even mild hypophosphataemia is associated with marke- dly increased mortality.
None.
Acknowledgements
We thank the staff and patients of Royal Victoria Hospital, Dundee. Dr Witham is supported by an NES/CSO Clinician Scientist award.
1 Medicine for the Elderly, Section of Ageing & Health, Ninewells Hospital & Medical School, NHS Tayside, Dundee DD1 9SY, UK
2 Department of Clinical Biochemistry, NHS Tayside, Dundee DD1 9SY, UK
3 Section of Ageing & Health, Department of Medicine & Therapeutics, University of Dundee, Scotland, UK
* To whom correspondence should be addressed Email: d.sumukadas{at}nhs.net
References
- Zazzo JF, Troche G, Ruel P, Maintenant J. High incidence of hypophosphatemia in surgical intensive care patients: efficacy of phosphorus therapy on myocardial function. Intensive Care Med (1995) 21:826–31.[CrossRef][Web of Science][Medline]
- Fisher J, Magid N, Kallman C, et al. Respiratory illness and hypophosphatemia. Chest (1983) 83:504–8.
[Abstract/Free Full Text] - Shor R, Halabe A, Rishver S, et al. Severe hypophosphatemia in sepsis as a mortality predictor. Ann Clin Lab Sci (2006) 36:67–72.
[Abstract/Free Full Text] - Kagansky N, Levy S, Koren-Morag N, Berger D, Knobler H. Hypophosphataemia in old patients is associated with the refeeding syndrome and reduced survival. J Intern Med (2005) 257:461–68.[CrossRef][Web of Science][Medline]
- Amanzadeh J, Reilly RF Jr. Hypophosphatemia: an evidence-based approach to its clinical consequences and management. Nat Clin Pract Nephrol (2006) 2:136–48.[Web of Science][Medline]
- Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis (1987) 40:373–83.[CrossRef][Web of Science][Medline]
- Hodkinson HM. Evaluation of a mental test score for assessment of mental impairment in the elderly. Age Ageing (1972) 1:233–38.
[Abstract/Free Full Text] - Folstein MF, Folstein SE, McHugh PR. Mini-mental state". A practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res (1975) 12:189–98.[CrossRef][Web of Science][Medline]
- Lum G. Hypomagnesemia in acute and chronic care patient populations. Am J Clin Pathol (1992) 97:827–30.[Web of Science][Medline]
- King AL, Sica DA, Miller G, Pierpaoli S. Severe hypophosphatemia in a general hospital population. South Med J (1987) 80:831–35.[CrossRef][Web of Science][Medline]
- Crook MA, Hally V, Panteli JV. The importance of the refeeding syndrome. Nutrition (2001) 17:632–37.[CrossRef][Web of Science][Medline]
- McWhirter JP, Pennington CR. Incidence and recognition of malnutrition in hospital. BMJ (1994) 308:945–48.
[Abstract/Free Full Text] - Nava S, Fuccella LM, Viglianti B. Physiological effects of intravenous fructose 1,6-diphosphate on diaphragmatic function in malnourished patients with COPD. Monaldi Arch Chest Dis (2004) 61:203–8.[Medline]
- Marchesani F, Valerio G, Dardes N, Viglianti B, Sanguinetti CM. Effect of intravenous fructose 1,6-diphosphate administration in malnourished chronic obstructive pulmonary disease patients with chronic respiratory failure. Respiration (2000) 67:177–82.[CrossRef][Web of Science][Medline]
- O’Connor LR, Wheeler WS, Bethune JE. Effect of hypophosphatemia on myocardial performance in man. N Engl J Med (1977) 297:901–3.[Abstract]
- Dhingra R, Sullivan LM, Fox CS JM, et al. Relations of serum phosphorus and calcium levels to the incidence of cardiovascular disease in the community. Arch Intern Med (2007) 167:879–5.
[Abstract/Free Full Text]
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||