Incidence of hypophosphataemia in patients on parenteral nutrition by oym20829


      For personal use only. Not to be reproduced without permission of the editor
  Incidence of hypophosphataemia
 in patients on parenteral nutrition
             By Vanessa Marvin, MSc, MRPharmS, Caroline May, BSc, Callum Livingstone, MRCPath, and Joy Davis, RD

      OBJECTIVE — To determine the effect of                        efeeding syndrome is a major            has been given safely in certain patient
      parenteral nutrition on plasma                                cause of hypophosphataemia in           groups have been published.12–14
      phosphate levels and to evaluate the                          hospital inpatients. It is defined as     A study was therefore undertaken in
      incidence of refeeding                                        “severe electrolyte and fluid shifts    which patients receiving PN at The Royal
      hypophosphataemia.                                associated with metabolic abnormalities in          Surrey County Hospital, Guildford (RSCH)
                                                        malnourished patients undergoing refeed-            were monitored in order to determine:
      SUBJECTS AND SETTING —                            ing, whether           orally, enterally       or
      250 adult patients started consecutively          parenterally”.1 After a period of fasting (the      ● The number of patients receiving PN
      on parenteral nutrition (PN) were                 length of which is not well defined, but              who develop refeeding
      monitored at The Royal Surrey County              might be as little as 48h),2 gluconeogenesis          hypophosphataemia
      Hospital, a 530-bed, non-teaching,                and protein catabolism are promoted and a           ● Whether any initial drop in plasma
      secondary care trust incorporating                negative protein balance occurs.When feed-            phosphate concentration in response to
      medical and surgical care, paediatrics,           ing with glucose is re-established (for               refeeding was sustained over the period
      intensive care, maternity services and a          example, when parenteral nutrition [PN] is            that patients received PN
      regional cancer centre.                           given) the cellular uptake of phosphate,            ● Whether there is any correlation
                                                        magnesium and potassium is promoted,                  between plasma magnesium and plasma
      DESIGN — Data on plasma levels of                 causing the concentration of these minerals           phosphate concentrations
      magnesium and phosphate, relevant                 in the plasma to decrease. Plasma concentra-
      interventions performed and nutrition             tions of thiamine are decreased and fluid and         Further aims of the study included the
      outcomes was collated by the Trust                sodium balance are also affected.                   identification of risk factors for developing
      Nutrition Support Team.                              Hypophosphataemia generally occurs               hypophosphataemia and the production of
                                                        after one to two days of PN in patients who         updated guidelines for medical staff at
      RESULTS — 36 patients (15 per cent)               were malnourished before being given PN,            RSCH on the biochemical tests required
      were found to be hypophosphataemic                and after three to five days in those who           before and during PN in specific patient
      before commencing PN. 86 patients                 were “normally nourished”.3 Hypomagne-              groups.
      (34.4 per cent) developed refeeding               saemia would be expected to accompany
      hypophosphataemia within seven days               phosphate depletion in malnourished                       Method
      after commencing PN. It was considered            patients,4 but it is not known whether mag-
      severe (less than 0.5mmol/L phosphate)            nesium levels reduce in line with phosphate         All adult patients who were started on PN
      in 10.8 per cent. Refeeding                       levels in those not significantly malnour-          during a 28-month period between 2003
      hypophosphataemia reached a nadir at              ished at the start of PN.                           and 2005 were invited to take part in the
      three days and recovered in over 75 per              Although hypophosphataemia has long              study. When first visited by the RSCH
      cent of patients within 10 days without           been recognised as perhaps the most impor-          nutritional support team, patients were
      intervention. Cancer patients appear to           tant aspect of refeeding syndrome,5 it is not       given information about the process of
      be at greater risk than non-cancer                always appreciated by prescribers of PN.6,7 It      delivering and monitoring PN and consent
      patients for developing refeeding                 would also appear that it cannot always be          to use their biochemical data in the study
      hypophosphataemia. Plasma                         prevented using best current practice and           and to hold the results (anonymised) on a
      magnesium levels did not mimic                    serious cases continue to be reported.8–10          database for future statistical analysis was
      phosphate levels.                                    It should also be noted that, when a plas-       requested. Providing consent was received,
                                                        ma sample is reported below the normal              patient data from medical notes, prescribing
      CONCLUSION — Recognition of                       range, before refeeding hypophosphataemia           and administration charts, dietetic records
      hypophosphataemia and the                         is diagnosed, other causes of hypophos-             and laboratory reports (see Panel 1, p167)
      development of guidelines for its                 phataemia need to be taken into account, as         were recorded by a member of the nutri-
      management in the care of PN patients             do the duration of the hypophosphataemia            tional support team onto the parenteral
      are important.                                    and the temporal association with starting          nutrition recording form and then entered
                                                        PN.                                                 anonymously onto an SPSS database (SPSS
                                                           We were unsure of the frequency of               version 14). Blood samples were collected
                                                        refeeding problems among the patients in            by members of the phlebotomy service and
                                                        our hospital receiving PN. Moreover, guid-          analysed by staff at the pathology laboratory
                                                        ance as to when to treat any fall in phosphate      service.The data contained in the laboratory
Vanessa Marvin is principal pharmacist,                 levels is lacking. There is also debate about       reports is set out in Panel 1 (p167). Local
Caroline May is chief aseptics technician, Callum       how much phosphate to give. Since 1994,             ethics committee approval for the study was
Livingstone is consultant clinical biochemist and       the British National Formulary has recom-           obtained.
Joy Davis is chief dietitian, all at The Royal Surrey
                                                        mended, giving 9mmol phosphate every                   In order to determine whether certain
County Hospital, Guildford. Correspondence to
                                                        12h, based on a small study.11 However,             patient groups were more prone to
                                                        reports of instances where a larger amount          hypophosphataemia, Pearson chi-squared

166    •                                                            H O S P I TA L P H A R M AC I S T                            M AY 2 0 0 7   •   VO L . 1 4
                                                 part in the study. The mean age was 63 years      concentrations actually increased after PN
                                                 (range 18–92 years). Almost half of patients      was started. One of these patients was devel-
     Panel 1: Data contained on                  (49 per cent) had undergone surgery during        oping acute renal failure, two received
         laboratory reports                      their stay, 44 per cent were diagnosed with       separate supplements of intravenous phos-
                                                 cancer, 8 per cent were diabetic and 40 per       phate and four patients improved with only
  Baseline blood levels (ie, before starting     cent were being treated in the hospital’s         the standard (18–30mmol per 24h) phos-
  parenteral nutrition [PN]) and blood levels    ICU for the course of their PN. Of the ICU        phate in their PN bag.
  during PN treatment of:                        patients, six were also on continuous
                                                 haemofiltration with phosphate-free dialysis      Incidence of refeeding hypophos-
  ■ Potassium, sodium, calcium,                  fluid.                                            phataemia A total of 86 patients (34.4 per
    magnesium, phosphate, urea,                                                                    cent) were deemed to have developed
    creatinine, alanine transaminase,            Baseline phosphate concentrations                 refeeding hypophosphataemia. In 13 of
    alkaline phosphatase and bilirubin           Plasma phosphate concentrations were              these patients, baseline phosphate levels had
    — all analysed daily for five days and       obtained in the 48h period before starting        been below the normal range, and fell fur-
    then three times a week.                     PN (ie, baseline phosphate levels) for 236        ther. All other patients had had phosphate
  ■ Albumin, C-reactive protein and              patients. These showed that 26 patients (11       levels above 0.8mmol/L. Further informa-
    triglycerides — all analysed weekly          per cent) were hyperphosphataemic (ie,            tion about the degree of hypo-
  ■ Glucose — analysed daily. Four-              plasma phosphate >1.4mmol/L). A total of          phosphataemia experienced by those
    hourly blood strip analysis was also         36 patients (15 per cent) were hypophos-          patients who developed refeeding hypo-
    carried out at first if necessary,           phataemic       (ie,    plasma      phosphate     phosphataemia is set out in Table 1.
    depending on insulin requirements.           <0.8mmol/L). All other patients had base-            The median for the plasma phosphate
                                                 line blood phosphate concentrations within        nadir was three days after starting PN (range
                                                 the reference range before starting PN.           one to seven days). It should be noted that,
analysis was performed for the following            Likely causes of baseline hyperphos-           for seven patients who had hypophos-
groups:                                          phataemia included renal failure, (12             phataemia after PN started, baseline
                                                 patients), parathyroid surgery (one patient)      phosphate levels were not available, and so
● Male patients, compared with female            and over-treatment with intravenous phos-         an assessment of whether these patients
  patients                                       phate (two patients). For the remaining           developed hypophosphataemia as a result of
● Patients on the intensive care unit,           patients, the recorded hyperphosphataemia         refeeding could not be made.
  compared with those on general wards           might have been an artefact, since a high
● Patients diagnosed with cancer,                level was found on only one occasion.             Recovery from hypophosphataemia
  compared with those who were not                  Causes of baseline hypophosphataemia           Details about the pattern of recovery from
● Patients who were receiving PN                 were thought to include receiving glucose         hypophosphataemia are listed in Table 2,
  following surgery, compared with those         infusion(s) (six patients), receiving insulin     p168.
  receiving it for another reason                infusion(s) (three patients), malabsorption          For 12 patients with moderate to severe
                                                 (three patients) and respiratory alkalosis (one   hypophosphataemia, their plasma phosphate
Phosphate concentrations: The phos-              patient).                                         concentration was still low after seven to 10
phate concentrations used in the study to                                                          days.In one of these patients,the plasma phos-
indicate the degree of phosphate depletion       Phosphate levels during PN In the first           phate concentration remained the same
were:                                            week after starting PN, serial plasma phos-       (0.56mmol/L) and, in the remainder, the con-
                                                 phate results were available for 235 of the       centrations rose but were still within the
● Reference range (ie, normal):                  236 patients for whom baseline phosphate          moderately hypophosphataemic range. None
  0.80–1.40mmol/L                                concentrations were available.A total of 126      of the patients with severe hypophos-
● Mild hypophosphataemia:                        patients had low blood phosphate concen-          phataemia remained severe. However in one
  0.65–0.79mmol/L                                trations recorded at some time between Day        patient with very severe hypophosphataemia,
● Moderate hypophosphataemia:                    1 and Day 7 of PN. The lowest recorded            follow-up plasma phosphate concentrations
  0.50–0.64mmol/L                                phosphate concentration was 0.17mmol/L            were not known because of the withdrawal of
● Severe hypophosphataemia:                      — this patient died less than seven days after    active treatment and investigations.
  0.30–0.49mmol/L                                PN was started. In 33 cases, the cause of
● Very severe hypophosphataemia: less            hypophosphataemia was not refeeding
  than 0.30mmol/L                                because there had been no period (48h or          Table 1: Incidence and degree of
                                                 more) of impaired nutrition before PN
Patients were deemed to have developed           started.                                          hypophosphataemia in patients
refeeding hypophosphataemia if they had             Of the 110 patients with cancer, 70 had           who developed refeeding
had a period of 48h or more without food         some degree of hypophosphataemia, com-
before PN was started and if a reduction of      pared with 58 of the 125 non-cancer
0.15mmol/L or more, to <0.80mmol/L,              patients. The Pearson chi-squared test indi-
was detected from their baseline phosphate       cated that a cancer diagnosis is a risk factor        Degree of               Number of
level in the first seven days of PN.The lowest   for hypophosphataemia in patients on PN           hypophosphataemia           patients (%)
phosphate concentration recorded during          (P=0.008), the relative risk being 1.46 (95
the first seven days of PN was used to deter-    per cent CI 1.09–1.95). No significant dif-       Mild                           23 (9.2)
mine the degree of hypophosphataemia.            ference in the incidence of hypo-                 Moderate                       36 (14.4)
                                                 phosphataemia was observed between                Severe                         25 (10.0)
                                                 ICU/general ward patients (P=0.338), sur-         Very severe                     2 (0.8)
                                                 gical/non-surgical patients (P=0.116) or
A total of 250 (136 male and 114 female)         male/female patients (P=0.120).                   The plasma phosphate concentrations that
adult inpatients were started on PN over the        In seven patients whose baseline phos-         correlate to the various degrees of
study period. All gave their consent to take     phate levels had been low, plasma phosphate       hypophosphataemia are indicated in the text

M AY 2 0 0 7   •   VO L . 1 4                               H O S P I TA L P H A R M AC I S T                                            •    167
                                                                                                        example, concomitant disease associated
                                                                                                        with reduced plasma phosphate concentra-
      Table 2: Pattern of recovery from hypophosphataemia on PN                                         tions. However, ICU patients are perhaps
                                                                                                        more likely to receive timely interventions
Degree of         Number of patients in whom plasma phosphate concentrations:                           to correct any low plasma phosphate levels
hypophosphataemia                                                                                       than those in general wards. Further analysis
(No of patients)     Recovered        Recovered      Were       Remained                                of data in the present study might be able to
                      without           after     unknown at      below                                 show whether this is the case and might
                  supplementation supplementation 7–10 days    0.80mmol/L                               identify other risk factors.

Mild (49)                        28                    0                    21                      0   Correlation between plasma magne-
Moderate (48)                    30                    1                    11                      6   sium      and      plasma      phosphate
Severe (27)                      17                    5                     0                      5   concentrations There was no clear link
Very severe (2)                   0                    0                     1                      1   between plasma magnesium and phosphate
                                                                                                        concentrations, suggesting that the hypo-
The plasma phosphate concentrations that correlate to the various degrees of hypophosphataemia          magnesaemia developed by some patients
are indicated in the text. Supplementation involved a patient receiving a phosphate infusion (50mmol    was caused by factors other than refeeding
over 24h). Patients who were not given separate supplementation received 18–30mmol per day of           syndrome.
phosphate from their parenteral nutrition (PN) alone.
                                                                                                        Guideline development Local practice
Magnesium levels Concentrations of                  phataemia in 63 per cent of surgical                guidelines had previously been developed to
plasma magnesium were found to be low (ie,          patients.17                                         try and prevent refeeding problems. These
below    the    reference    range      of             When considering severe (and very                advocated the monitoring of electrolytes,
0.7–1.0mmol/L) in 30 of the 79 patients             severe) hypophosphataemia only, other stud-         glucose and circulatory volume before and
who developed refeeding hypophos-                   ies have shown an incidence of between              during refeeding and, where possible, cor-
phataemia and whose magnesium levels                0.97 per cent (<0.32mmol/L)2 and 1.7 per            recting any abnormalities before starting PN,
were measured. The following should be              cent (<0.4mmol/L).16                                (which can usually be achieved within 12–24
noted:                                                 The wide variation in the incidence of           hours).1 The problem with this approach is
                                                    hypophosphataemia most likely results from          that deficits of potassium, magnesium and
● Plasma magnesium concentrations of 10             a combination of factors.These include:             phosphate may not be corrected adequately
  of these 30 patients were low before PN                                                               without commencing some intake of calo-
  was started                                       ● Differing reference ranges used to                ries, and patients remain at risk of refeeding
● Plasma magnesium concentrations in                  define hypophosphataemia                          syndrome even if their plasma levels appear
  the two patients who developed                    ● Inclusion of patients fed enterally as well       “normal”. Another issue is that the fluid and
  refeeding hypophosphataemia and                     as parenterally                                   sodium load necessary to deliver potassium
  developed very severe                             ● Inclusion of groups of patients who               and phosphate can add to the patient’s risk of
  hypophosphataemia were normal                       have putative risk factors for developing         developing cardiac symptoms with refeeding
  throughout the study                                hypophosphataemia, such as cancer                 syndrome.The National Institute for Health
● In total, plasma magnesium                        ● Inclusion of patients who have                    and Clinical Excellence advocates starting
  concentrations fell in 20 of the 86                 established risk factors for developing           PN as soon as possible and not “correcting”
  patients (23 per cent) who developed                hypophosphataemia (eg, those taking               abnormal biochemistry first.18 This advice,
  hypophosphataemia as a result of                    medicines such as antacids,                       together with the findings from this study
  refeeding                                           bisphosphonates, glucocorticoids,                 that hypophosphataemia is not generally
● Of the 33 patients who had                          insulin and some antivirals)                      prolonged and that plasma phosphate con-
  hypophosphataemia after starting PN,              ● Selection bias (eg, studies that rely on a        centrations rise to normal in most patients
  but which was deemed not to be                      retrospective analysis of laboratory data         on PN without intervention, have been
  refeeding hypophosphataemia, eight                  might be expected to find a higher                taken into account in updating the local
  patients (24 per cent) also had                     incidence of hypophosphataemia                    guidelines. Other changes made to guide-
  hypomagnesaemia                                     because patients with a history or risk of        lines and policies include:
                                                      developing the condition are more
      Discussion                                      likely to have blood samples taken for            ● The section in the local clinical
                                                      biochemistry analysis and therefore be              guidelines booklet on PN has been
Incidence of hypophosphataemia and                    included).                                          updated to improve the ease in which
refeeding hypophosphataemia Previous                ● Differing amounts of phosphate                      biochemical tests are ordered
studies have shown a wide range of inci-              consumed or given before PN is started            ● A section of the guidelines on the
dence of refeeding hypophosphataemia in             ● Differing amounts of phosphate                      treatment of hypophosphataemia has
patients receiving nutritional support. For           contained in PN                                     been prepared
example, in a study of 158 patients who                                                                 ● A PN policy has been agreed by the
received either enteral or parenteral nutri-        Risk factors for developing hypo-                     RSCH NHS trust which contains
tion following an intestinal fistula, the           phosphataemia We found that there was                 sections on refeeding syndrome
incidence was found to be 9.5 per cent.15           an increased risk of hypophosphataemia in
Another study showed that moderate to               cancer patients.We are unaware of any other           It should be noted that a complication in
severe hypophosphataemia occurred in 34             published research about this in PN patients.       developing guidelines is whether to take
per cent of 62 ICU patients in a US hospital          In our study, we did not find that patients       account of the phosphate contained within
(reference range 0.84-1.58).2 In addition, a        on ICU were any more prone to hypophos-             the lipid in PN (typically 6–7.5mmol). Con-
Spanish research group found that the inci-         phataemia than patients on general wards.           ventionally, the phosphate content of PN,
dence of hypophosphataemia was 16 per               This is possibly surprising, given that ICU         including that within lipids, is considered to
cent.16 French researchers found hypophos-          patients might be more likely to have, for          be 100 per cent bioavailable.19 However,

168   •                                                         H O S P I TA L P H A R M AC I S T                            M AY 2 0 0 7   •   VO L . 1 4
there is little evidence that patients,                   2. Marik PE, Bedigian MK. Refeeding
especially the critically ill, can use the lipid              hypophosphataemia in critically ill patients in an
phosphate, and so it is difficult to know                     intensive care unit. A prospective study. Archives of
whether to include it in calculations. Either                 Surgery 1996;131:1043–7.
way, additional phosphate (ie, other than that            3. Brooks MJ, Melnik G. The refeeding syndrome: an
within lipids) is necessary to provide the rec-               approach to understanding its complications and
ommended daily amount (of 17.5mmol)                           preventing its occurrence. Pharmacotherapy
and there is typically a total of 18 to 30                    1995;15:713–26.
mmol phosphate in the majority of modern                  4. Weisinger JR, Bellorin-Font E. Magnesium and
compounded solutions.                                         phosphorus. Lancet 1998;352:391–6.
                                                          5. Ritz E. Acute hypophosphataemia. Kidney
Limitations This study was performed in a                     International 1982;22:84–94.
heterogenous group of adult patients, most                6. Marinella MA. The refeeding syndrome and
of whom were receiving PN following                           hypophosphataemia. Nutrition Reviews
surgery.Although the findings are generalis-                  2003;61:320–3.
able, it might be difficult to link risk factors          7. Hearing SD. Refeeding syndrome. BMJ
to causation.                                                 2004;328:908–9.
   Blood phosphate concentrations are a                   8. Duerksen DR, Papineau N. Electrolyte abnormalities
measure of total plasma inorganic phos-                       in patients with chronic renal failure receiving
phate, which generally accounts for just 0.1                  parenteral nutrition. Journal of Parenteral and
per cent of total body phosphate or less.The                  Enteral Nutrition 1998;22:102–4.
concentrations therefore do not necessarily               9. Shadaba A, Paine J, Adlard R, Dilkes M. Re-feeding
correlate with the development of clinical                    syndrome. Journal of Laryngology and Otology
symptoms, until levels drop significantly.                    2001;115:755–6.
They are, nevertheless, seen as a useful mark-            10. Slodkowski M, Rubinsztajn R, Cebulski W,
er of phosphate depletion.                                    Krasnodebski-Ireneusz W. A case report of severe
   In addition, there is no real national con-                hypophosphataemia in the course of refeeding
sensus about what concentration of plasma                     syndrome. Polski Merkuriusz Lekarski
phosphate constitutes, for example, severe                    2004;17(102):638–9.
hypophosphataemia. For our study, we                      11. Vannatta JB, Whang R, Papper S. Efficacy of
therefore adapted parameters used in other                    intravenous phosphorus therapy in the severely
studies.14 Had we chosen different parame-                    hypophosphatemic patient. Archives of Internal
ters, some of our results might have been                     Medicine 1981;141:885–7.
different.                                                12. Terlevich A, Hearing SD, Woltersdorf WW, Smyth C,
                                                              Reid D, McCullagh E et al. Refeeding syndrome:
Future work Further analysis of the data                      effective and safe treatment with Phosphates
has been carried out to determine whether                     Polyfusor. Alimentary Pharmacology and
any risk factors could be determinant for                     Therapeutics 2003;17:1325–9.
refeeding syndrome, including the amounts                 13. Perreault MM, Ostrop NJ, Tierney MG. Efficacy and
of phosphate given in PN. It is hoped that                    safety of intravenous phosphate replacement in
analysis of the data will show how long a                     critically ill patients. Annals of Pharmacotherapy
period of undernutrition or “nil by mouth”                    1997;31:683–8.
is determinant for refeeding syndrome.                    14. Charron T, Bernard F, Skrobik Y, Simoneau N,
                                                              Gagnon N, Leblanc M. Intravenous phosphate in the
        Conclusion                                            intensive care unit: more aggressive repletion
                                                              regimens for moderate and severe
All 250 patients on PN over 28 months                         hypophosphataemia. Intensive Care Medicine
were recruited to the study and relevant bio-                 2003;29:1273–8.
chemistry results were available for almost               15. Fan CG, Ren JA, Wang XB, Li JS. Refeeding syndrome
90 per cent of patients. Over a third of                      in patients with gastrointestinal fistula. Nutrition
patients receiving PN developed refeeding                     2004;20:346–50.
hypophosphataemia, which was severe in                    16. Llop Talaveron JM, Comas Sugranes D, Badia Tahull
10.8 per cent of cases.The nadir was found                    MB, Saez Fernandez A, Jodar Masanes R, Gomaz
to be Day 3. Most patients did not require                    Saez JM. hypophosphataemia in parenteral
intervention to resolve their low phosphate                   nutrition: prevention and associated risk factors.
levels because they rose during the next                      Nutricion Hospitalaria 2004;19:362–6.
seven days with continuing PN. Cancer                     17. Zazzo JF, Troche G, Ruel P, Maintenant J. High
patients on PN may be more at risk of                         incidence of hypophosphataemia in surgical
hypophosphataemia than others.                                intensive care patients: efficacy of phosphorus
   Falls in magnesium levels mimicked falls                   therapy on myocardial function. Intensive Care
in phosphate levels in less than a quarter of                 Medicine 1995;21:826–31.
identified cases of refeeding hypophos-                   18. National Institute for Health and Clinical Excellence.
phataemia.                                                    Nutrition support in adults. Oral nutrition support,
                                                              enteral tube feeding and parenteral nutrition.
        References                                            Clinical Guideline. London: NICE; Feb 2006.
                                                          19. Pineault M. TPN in very low birth weight infants
1. Crook MA, Hally V, Panteli JV. The importance of the       with Travasol blend C. Journal of Parenteral and
   refeeding syndrome. Nutrition 2001;17:632–7.               Enteral Nutrition 1986;10:296–9.

M AY 2 0 0 7   •   VO L . 1 4                                           H O S P I TA L P H A R M AC I S T

To top