Postscript to: 'A pragmatic approach to managing severe malnutrition: Is F75 always beneficial?'

Michael Golden,Yvonne Grellety

It is quite wrong to consider the advantage of F75 as "theoretical". However, a decreased mortality will not be seen if other aspects of faulty management dominate mortality. It is important that readers of Field Exchange understand the reasons and the underlying physiological changes that occur in conjunction with malnutrition, because they have much wider implications for management than simply whether or not to use F75. It is now clear that if fluid and electrolytes (including those contained in the diet) are not managed correctly there will be a high mortality.

The metabolism of sodium and the function of the kidney and heart are very abnormal in all malnourished patients. They are very sensitive to excess of sodium intake from any source. It is the total intake that matters. The diet only forms part of this sodium load and any benefit from using a low sodium diet is immediately removed when excess sodium comes from other sources (resomal, infusion, ORS, transfusion).

The rate at which sodium and other electrolytes are filtered in the kidneys is only about half normal in a malnourished child. If given an infusion of sodium the malnourished child retains most of the sodium in the body. If given ReSomal the amount of fluid that the well nourished child with diarrhoea can excrete in 20 minutes takes 10 hours in the malnourished child! The well nourished child with diarrhoea can excrete the excess rapidly with no danger of fluid overload. In contrast the malnourished child with changes in kidney function will retain more and more until the patient either gets peripheral oedema, acute pulmonary oedema or heart failure. The situation is even more critical in children with kwashiorkor.

How much sodium is too much? Data from Uganda showed that 17% of children developed heart failure when the total sodium intake was above 3 mmol sodium/ kg/d. Few children went into heart failure when the intake was less than 1 mmol/kg/d of sodium. Resomal has about 45 mmol sodium per litre. F100 has 19 mmol sodium/ litre (1000kcal). F75 has 6 mmol/ litre (8 mmol/ 1000kcal). Thus, an unacceptable number will get heart failure if more than 70 ml/kg of resomal, 160 ml/kg of F100 or 500 ml/kg of F75 is given (these are the amounts that give 3 mmol/kg sodium) in excess of losses. With Resomal and F100 these quantities are easily and regularly exceeded.

The situation is even more finely balanced during early recovery. Inside the marasmic child's cells the average sodium concentration is about 70 mmol/ litre cell water whereas the normal child has 35 mmol/ litre cell water. Thus, during early recovery 35 mmol sodium /litre of cells has to come out of the cells into the circulation and be excreted. If this happens before the child's kidneys recover the patient may go into heart failure. The sodium coming out of the cells will either be deposited in the interstitial space giving "refeeding oedema" or be excreted by the kidney or remain in the plasma. For a typical malnourished child weighing 10 kg, if the sodium is retained in the circulation the plasma volume will increase by 16% per day and the haematocrit will fall by about 3% per day. Even larger falls in haematocit are frequently seen in practice. Even without any additional sodium intake there is the potential for a major increase in plasma volume simply from reversal of the electrolyte imbalance. If large amounts of F100 are given during recovery this further increases the risks of an expanded circulation.

These phenomena have been known for a very long time. Researchers working in therapeutic feeding centres in Jamaica strictly limited sodium intake in malnourished children. The mortality was about one fifth of that expected and for one period of 18 months there were no deaths. In Tubmanbourg, Liberia where no infusions were used and glucose-water was used instead of sodium containing fluids, there were only 16 deaths from 900 patients which again was much less than expected. There have been similar findings in Burundi, Bangladesh and Somalia.

A large number of children have already died to give the data and insights used to design F75 and the other aspects of modern protocols. The same applies to adults - in experiments in Minnesota during the war 4 of 13 patients who had their central venous pressure measured were found to be in incipient heart failure during refeeding and one patient (no 47) went into overt heart failure.

Recently we have analysed the outcome of about 8500 children comparing the observed to expected mortality rate. It is clear from this analysis that those centres where children have a high initial weight gain (reflecting fluid retention) have an excessive mortality and that most excess deaths are preceded by at least one days weight gain. It is also clear that where there is no appropriate control of other aspects of fluid and electrolyte management or too rapid progression through the phases, then no advantage can be shown to come from the use of F75. Indeed, because the rate of recovery and the cellular electrolyte changes are much slower with HEM than with F100, there can be a lower mortality with HEM, than with F100 itself if the F100 is not used with understanding. These diets are strong drugs and as with all drugs should only be used by those with appropriate training.

There are several common ways in which patients get fluid overload, heart failure and die. First, they are given excess fluid at admission because of confusion over the signs of dehydration and malnutrition (oedematous children by definition are already "overhydrated" although they may have low circulating volume). Second, the child is started on F100 and the amount given increased too quickly. He then gets an osmotic diarrhoea which is then treated with Resomal, instead of simply changing the diet to remove the intestinal osmolar load. Third, the child is noted to have a fall in haemoglobin during treatment and this is "treated" with a transfusion. The haemoglobin ALWAYS drops as the sodium comes out of both the cells and the interstitial spaces and increases the plasma volume without increasing the red-cell volume - indeed the extent of fall of the haemoglobin can be used as a measure of the degree of expansion of the plasma volume and the additional stress the heart is under. A large fall in haemoglobin is a real danger sign for heart failure and must not be treated with a transfusion that further expands the circulation at this critical time. Indeed, many of these children are diagnosed as dying from "anaemia".

How does this translate into practical procedures at the field level?

1. Oedematous patients should not be given infusions or resomal unless they are unconscious from hypovolaemia, and then the amounts should be very small. Resomal should, in general not be used in oedematous children.
2. Patients should be weighed daily, after admission if they have not lost weight, even if they have diarrhoea, they should not be treated for dehydration, for weight change is by far the best measure of net fluid balance.
3. Diarrhoea, during treatment, should be treated with a change of diet - amount and/or quality (back to F75).
4. The liver edge should be marked on admission with an indelible pen. If the liver size increases by more than 2 cm then the patient's circulation is increasing and there is incipient heart failure. Those who have an increased respiratory rate with a steady or increased weight have heart failure (pneumonia gives rise to weight loss).
5. If anaemia is to be treated by transfusion this must be within 48h of admission and no food or fluid should be given that day. The very maximum that should be given is 10 ml whole blood per kilo over at least three hours with close surveillance. After this time Hb can be used as a measure of heart failure (and if very severe anaemia has to be treated this should only be done with a small exchange transfusion - just as in a neonate).
6. Intravenous fluids should not be given. Administration of quinine or antibiotics should not be by infusion of a sodium containing solution. When the child's haemoglobin drops and he dies the wrong diagnosis will be made.
7. All complications in phase two should be first managed by putting the child back to phase 1.
8. The introduction of a transition phase between phase one and two has decreased the mortality from overfeeding in the initial phases of treatment.
9. F75 should be used along with the other ways in which the total sodium intake is restricted.

The physiology is clear with respect to sodium homeostasis (control). However, the use of F75 will not solve many of the problems seen in a centre run by inappropriately trained or supervised staff, where the benefits gained from F75 are overshadowed by the detrimental effects of other forms of faulty management. On the other hand giving F75 is exceptionally simple to implement and will do no harm. It is certainly much easier and less likely to mishap than many other procedures that are seen to be "essential" skills in many TFCs, such as diagnosing and treating malaria, putting up a drip, cross-matching and giving blood, or even maintaining a cold chain to give a vaccine.

For further information contact Michael Golden at mikegolden@eircom.net or Yvonne Grellety at ygrellety@wandoo.fr.

View the article that this postscript relates to