Anaemia in Kwashiorkor
Anaemia in Kwashiorkor Fitsum Assefa 22.01.99
RE: Anaemia in Kwashiorkor steve collins 22.01.99
iron witholding in kwashiorkor Kay Tomashek 23.01.99
re withholding iron in kwashiorkor Michael Golden 23.01.99
Re: Anaemia in Kwashiorkor Nevin Scrimshaw 23.01.99
Re: Anaemia in Kwashiorkor Nevin Scrimshaw 23.01.99
iron and KWK André Briend 23.01.99
Re: iron and KWK-dose Michael Golden 25.01.99
Re: Iron supplementation in Malnourished children(Kwash) Andre Renzaho 27.01.99
Re: iron and KWK Andre Renzaho 27.01.99
RE: camp data Steve Hansch 26.01.99
climate and treatment George Beaton 27.01.99

Date: Fri, 22 Jan 1999 13:51:38 +0000

From: "Fitsum Assefa" <>

Subject: Anaemia in Kwashiorkor


Dear Mike,

I have a question about anaemia in severely malnourished, especially Kwashiorkor, adults.

Since last August, Burundian refugees, have been arriving at Kigoma port, most of them (over 70% ) are young men. They left their homes in Burundi more than 1 year ago and have been running from one bush to another, until they finally got a chance to escape. During this time they mainly depended on raw Cassava, some wild leaves /fruits and banana. They were not able to cook any of their food because the smoke would direct the army to their hiding place.

We have observed unusually high rates of severely malnourished adults (mostly men), especially Kwashiorkor; they have a very high rate of anaemia, many very severe anaemia, Hb <4g/100ml. There is a particularly high mortality associated with the anaemia. The district hospital in Kigoma do blood transfusion but this is mostly unsuccessful (They check blood for HIV, but not for Hepatitis).

Other than the very sever anaemia, almost everybody admitted in our TFC has visible anaemia. We are following the proper TFC protocol in our centre, using F75/100, withholding iron for the initial stage of treatment, giving folic acid from day 1 etc, etc..

I know that the aetiology of anaemia is usually multi-factorial. Malaria is endemic and I expect it is one of the important contributory factors to the anaemia.

The anaemia often gets worse during the nutrition rehabilitation period (up to about 2 weeks). Cases who had visible anaemia on admission can require transfusion, and some who did not have visible anaemia developed visible anaemia.

Many health staff in the area have the opinion that the anaemia is due to iron deficiency, and want to give Fe supplementation at an early stage; the district hospital used to give Fe to all severely malnourished, including Kwashiorkor cases, from day one.

I think there has not been enough investigations made to determine the cause/s of the anaemia among this group or the correct treatment.

What is your advice?


Fitsum Assefa.

From: "steve collins" <>

Subject: Ngonut: RE: Anaemia in Kwashiorkor

Date: Fri, 22 Jan 1999 17:20:19 -0000



When you say that these adults are severely malnourished are you basing that on the fact that they have oedema or are they very wasted as well. (i.e. what are their MUACs?) I ask because it appears to me that in the adults I have seen with oedema the oedema can have different prognostic significance depending upon the degree of wasting and complicating clinical illness also present. If the patients are also very wasted ( i.e MUAC around the 16 - 18.5 cm mark or there are other signs of liver failure (petchial haemorrhage / jaundice)) the oedema marks a very bad prognosis (especially in men) and I would treat them as per the phase one guidelines with no iron and F75 and routiene antibiotics, but starting their calorie intake low at around 40 - 50 Kcal / Kg / day (c.f. the 75 - 100 for children). If on the other hand they are not so wasted (according to MUAC) in my experience I feel that the anaemia (especially if at 4 gm HB) is much more likely to kill them than some Fe or too much F100. With Hbs of 4 gm are you sure that it's not high output heart failure that you are seeing?




Dr Steve Collins

Oleuffynon, Old Hall, Llanidloes

Powys, SY18 6PJ, Wales, UK.

tel: 44 (0) 1686 413989, fax: 44 (0) 870 1641364,

Date: Sat, 23 Jan 1999 16:16:57 +0000

From: "Tomashek, Kay" <>

Subject: Ngonut: iron witholding in kwashiorkor


I have been following the NGO-nut messages from Fitsum Assefa and Steve Collins concerning iron therapy in TFC patients. I am wondering if you know where the TFC protocol recommendation of waiting two weeks to begin iron therapy originated. TFC protocols do not give the study references since the practice is so well established. In a brief literature search I could only find rat studies from the 70's (Olusi SO, McFarlane H, 1978) that support the notion that iron supplementation during the treatment of PEM may predispose to bacterial infection. I could not find any clinical trial in humans that has investigated this finding. Thanks for your assistance.


Kay Tomashek


Kay M. Tomashek, MD MPH, EIS Officer

Centers for Disease Control and Prevention (CDC), International Emergency and Refugee Health Branch, Mailstop F-48, 4770 Buford Hwy, NE, Chamblee, Georgia 30341 USA

Phone: (770) 488-4466, Fax: (770) 488-7829, Email:

Date: Sat, 23 Jan 1999 16:16:35 +0000

From: Michael Golden <>

Subject: Ngonut: re withholding iron in kwashiorkor.


Dear Kay,

In the early 1980's, in Jamaica, after we were convinced that protein deficiency was not the cause of kwashiorkor, we decided to re-look at that role of ferritin because of the work of Srikantia using functional bioassays. We confirmed that ferritin was high (Ramdath and Golden), that urinary iron excretion after desferrioximine was also high and that the livers of Jamaican children dying from malnutrition had very high iron contents, and that the various measures of iron overload were closely related to mortality. Gilman & Gilman (1951) had also described hepatic iron overload in South Africa as had Waterlow in Jamaica in 1948. The high iron in association with the low glutathione content of their red cells on admission and was the starting point for the free-radical hypothesis (Golden and Ramdath).

It had been known for many years that the transferrin level was very low in kwashiorkor, and we calculated that many children would have free iron in their plasma (Ramdath and Golden) - since confirmed by direct measurements in South Africa (Dempster). It is well known from the toxicological literature on iron poisoning in children that free iron in the plasma is associated with a high mortality.

When we were doing this work Franca Smith (1989), from Nigeria, was doing a sabbatical with us in Jamaica. She then examined her data from Ife Ife to see if there was a relationship between iron therapy and mortality - the hospital in Ife frequently ran out of drugs at that time - what she found was that the mortality was substantially higher in her ward when the pharmacy had iron and it was given to the children than when the pharmacy ran out of iron. We analysed the data from the children whose samples she had brought from Nigeria to Jamaica and it showed the same thing - although it just failed to reach statistical significance.

With the strong association between measures of iron status and death in Jamaica, the demonstration of iron overload and free iron in these children's plasma and the "natural" experiment in Nigeria, we judged that it would not be ethical to conduct a formal clinical trial of "with and without iron", and started to strongly advocate that iron should not be given during the early stages of treatment.

We now have examined the mortality data from Jamaica in relation to the prognostic index developed by Claudine Prudhon and find that in the children treated without any iron that the observed over expected mortality for marasmus is about 20% and for kwashiorkor is just under 50%. I suggest that we can say from these data that withholding iron does not jeopardise the children at all.

The next question to be answered is: for how long should iron be withheld?

I would answer,

1) Certainly until the liver has recovered to the stage where it has synthesised and secreted sufficient transferrin for additional administered iron not to pose a direct threat of saturating the this critical transport protein. If it is given before this then a. there is a danger of direct iron poisoning, b. iron will be made available for invasive bacteria, and c. it cannot be transported and used for incorporation into haem at any rate so it will put the patient in jeopardy for no possible therapeutic benefit.

The resynthesis of transferrin tends to happen quite rapidly and it is usually at acceptable levels in about one week, in children who regain their appetites and loose their oedema/start to gain weight. Although the time course of recovery of transferrin has not been adequately studied in relation to clinical parameters (that have to be used in nearly all facilities treating severe malnutrition).


2) There is another consideration with kwashiorkor though. In our Jamaican children it takes about 2 weeks from the intracellular sodium to come down to normal levels and the potassium to rise (Alleyne GAO). It also takes two weeks for the red cell glutathione to increase to normal levels and for the red-cell NADPH/NADP ratio to become normal (Ramdath - unpublished). I have measured urinary nitrate excretion on a nitrate free diet (measure of nitric oxide production) and it remains high for much longer than I had anticipated clinically (until after loss of oedema and start of rapid weight gain). Lastly, Bernabeau (unpublished) was measuring the rate of height change in marasmic and kwash children in Sierre Leone with ACF. Interestingly, there was an almost immediate commencement of height gain in the marasmic children - but the height gain did not start for 2 weeks in the kwashiorkor children. These various studies show that the metabolic defect, oxidative stress and electrolyte imbalance are not corrected in kwashiorkor for about 2 weeks after the start of treatment, even if the children are getting on quite well clinically. I think it would be prudent not to give these patients an iron stress until such time as they have corrected the underlying metabolic defect. So at the moment I am advocating, for kwashiorkor that iron should be withheld for 2 weeks after the start of treatment.

It is often not feasible in a refugee camp/TFC situation to stick strictly to this time frame as all children who are rapidly growing are given the same diet (the iron is incorporated into the diet at this stage for logistical reasons so that each child does not need to be medicated by the staff each day). There does seem, in some situations, to be an excess of deaths (observed over expected), from days 5 to 14 and we are investigating this at the moment - it is perfectly possible that it is due introduction of iron to early.

On the other hand, it would not be correct to withhold iron during the middle/later stages of recovery, when the patients have repaired their iron metabolising machinery, are gaining weight rapidly, having to increase their circulating red cell mass with their body mass and make good any anaemia that they have.

I do think there is room for a clinical trial at this stage on early (say day 7-14), later introduction (say day 14 to 21), or very late (day 28+) of iron treatment.

Best wishes,


Mike Golden.

Date: Sat, 23 Jan 1999 13:59:10 +0000

From: "Dr. Nevin Scrimshaw" <>

Subject: Re: Ngonut: Anaemia in Kwashiorkor


Replying to Dr. Assefa's question takes me back a long ways. However, it is worth remembering that the protein deficiency of Kwashiorkor causes a mild normocytic anemia due to deficient protein synthesis but superimposed on this may be severe micronutrient deficiency of iron deficiency, severe macrocytic anemia of folate deficiency and other B-vitamins may also be limiting erythropoiesis.

When we were studying this at INCAP in the 1950's when marasmic kwashiorkor was abundant initial treatment with protein and energy alone brought a mild short-lived reticulocyte response. In most cases giving iron then resulted in a marked and prolonged recticulocyte response which usually ceased without fully correcting the anemia. With folate and other B-vitamins a third reticulocyte response occurred and hemoglobin eventually returned to normal;

We noted that the nature of the superimposed anemia varied with the origin of the child. Those from the lowlands of Guatemala, generally had hookworm infection, severe iron deficiency and a microcytic hypochromic anemia. Many cases from El Salvador had a macrocytic anemia superimposed which we believed was du to primary folate deficiency. As you can imagin in most case the etiology was multiple and sometimes reflected in microcytic and macrocytic cells in the same blood smear.

I think if probable that the situation of multiple etiologies is similar in your refugee situation. Any sustained response will depend not only on correcting the protein deficiency relative to energy intake, but also supplying both iron and B-viitamins. The only caution is to avoid large doses of iron so as not to overwhelm the ability of the reduced transferrin levels to withhold iron from pathogenic organisms while immune mechanisms are recovering from the malnutrition and transferriin levels have returned to normal.


Nevin Scrimshaw, UNU

Date: Sat, 23 Jan 1999 21:07:54 +0000

From: "Dr. Nevin Scrimshaw" <>

Subject: Re: Ngonut: iron withholding in kwashiorkor


Perhaps the following excerpt from an old article of mine will help Kay Tomashek with references.

Nevin Scrimshaw



Nevin S. Scrimshaw, Institute Professor Emeritus

Massachusetts Institute of Technology, Cambridge


Director, International Food and Nutrition Programme

United Nations University, Tokyo

Published in the "Report of the Third Annual Nutrition Workshop Meharry

Medical College - Center for Nutrition, 1989


Overload and infection

It is important for this workshop to discuss the consequences of excessive dietary iron as well as iron deficiency for two reasons. First, there are circumstances in which the administration of parenteral or large oral doses of iron can lead to overwhelming infections. Second, failure to understand the special circumstances under which this occurs can interfere with needed programs of iron supplementation and fortification for populations at risk of iron deficiency.

It is not only the host that needs iron for the biochemical functions mentioned in the introduction, but also the infectious agent (66). Without it replication is inhibited. In fact, withholding iron from the infectious agent appears to be an important resistance mechanism of resistance to infections (67, 68). Conalbumin and lactoferrin have stronger iron binding properties than most bacterial siderophores and are normally highly unsaturated and function as iron-withholding rather than iron-transport agents.

Lactoferrin, known to be released upon degranulation of leucocytes in aseptic areas, is a major component of human milk and resistsproteolytic destruction in the gastrointestinal tract. It is not difficult to demonstrate in vitro the protective effect of lactoferrin in vitro (69). In the iron- deficient host with reduced cell mediated immunity and leukocyte function, lack of available iron for agent replication is protective. Baggs and Miller (24) found that in rats exposed to a standard dose of Salmonella, a diet lacking in iron is almost as protective as one meeting iron needs.

Murray and Murray have described the exacerbation of malaria with a high case fatality rate in Somalia refugees given therapeutic doses of iron (70, 71). Parenteral iron given to children with kwashiorkor has also been associated with mortality from overwhelming infection (72, 73).

However, in our supplementation studies in Egypt (29) and Indonesia (31) morbidity from diarrheal and respiratory disease decreased promptly with iron supplementation. The iron supplement promoted recovery of immune function without providing enough iron to increase the severity of existing or subsequent infections.


66. Dillman RB, Mackler D, Johnson C, -Brengelrnann W, Green C, Gale J, Martin, M. Layrisse, C Martinez-Torres, and CA Finch. Effect of iron deficiency on catecholamine metabolism and body tempera-ture regulation. In Iron Deficiency: Brain Biochemistry and Behavior, New York: Raven Press, 1982.

67. Sussman M. Iron and infection. Chapter 19 in Iron in biochemistry and medicine, 649-679. London and New York: Academic Press, 1974.

68. Rogers H. Bacterial iron metabolism and host resistance. Microbiology:289-322, 1974.

69. Weinberg ED. Infection and iron metabolism. Am J Clin Nutr 30:1485-1490, 1977.

70. Bullen JJ. Acute diarrhoea in childhood. Ciba Foundation Symposium, 149-162. Amsterdam: Elsevier/North Holland, 1976.

71. Murray MJ, Murray AB, Murray BA, Murray MB. Diet and cerebral malaria: The effect of famine and refeeding. Am J Clin Nutr 31:57-61, 1978.

72. Murray MJ, Murray AB, Murray MB, Murray CJ. The adverse effect of iron repletion on the course of certain infections. Br Med J 2: 1113-1115, 1978.

73. McFarlane HS, Reddy KJ, Adcock, Adeshina H, AR Cooke and J. Akene. Immunity, transferrin and survival in Kwashiorkor. Br Med J 4: 268-270, 1970.


29. Hussein MA, Hassan HA, Abdel-Chaffar AA, Samem S. Effect of iron supplements on the occurrence of diarrhoea among children in rural Egypt. Food and Nutr Bull. 10(2):35,1989.

31. Hussaini MA. The use of fortified salt to control vitamin A deficiency. Bogor Agricultural University, Bogor, Indonesia, 1982.

Date: Mon, 25 Jan 1999 10:18:00 +0100 (CET)

From: (Andre' BRIEND)

Subject: Ngonut: iron and KWK


Dear all,

I think Mike made a strong case against iron in severe malnutrition.

However, I feel a key piece of info is missing in the Smith et al paper (Eur J Clin Nutr 1989; 43: 763-8) which tilted the balance of evidence against early iron use: there is no mention in the paper of the dose used in this study associated with increased mortality.

Mike do you know about the iron doses given in this study ? Was it a high or a moderate supplement ?

Maybe there is room for studies with and without moderate levels of iron supplement along with studies you mention about the timing of introduction of iron.



Dr. André Briend

Groupe Nutrition Santé ISTNA, 5 rue du Vertbois, 75003 Paris, France

tel : 33-1-53 01 80 36, fax : 33-1-53 01 80 38

Date: Tue, 26 Jan 1999 18:20:13 +0000

From: Michael Golden <>

Subject: Re: iron and KWK-dose


Dear André

I have the relevant file in Ireland not here in Aberdeen so I cannot answer your question unequivocally at the moment - I know the iron was given once per day and the weight of the child was not taken into account - from memory, the dose was 40mg elemental iron per child per day.

Thus, a 10kg child would have had about 4 mg/kg/d, not an excessive dose by any means.



Prof. Michael H.N.Golden


Date: Wed, 27 Jan 1999 11:13:08 +1000 (EST)

Subject: Re: Iron supplementation in Malnourished children(Kwash)




Iron supplementation in early treatment of Kwashiorkor may lead to production of free radicals.Acytivated oxygen species are formed during the stepwise reduction of oxygen to water, and by secondary reactions with protons and transition metals such as copper and iron. These highly activated oxygen species eg superoxides are removed from the body by anti-oxydants such as Vit A and Zinc. However malnourished children do not have this capacity. Therefore, in malnourished children, these free radicals can build up enough to damage tissues leading to oedema, fatty enlargement of the liver, pale skin and hair, and possibly diarrhoea.

1. Additional readings: Garrow J.S and James W.P.T(1994), Human nutrition and dietetics, churchill livingstone, p .225

2. King F.S and Burgess A. (1992)Nutrition for developing countries, 2nd edn, Oxford university press,p. 223.


Andre M.N. RENZAHO, Nutritionist and dietician

MPH-Health program evaluation (currently completing)

Deakin University, Melbourne/ Australia


Date: Wed, 27 Jan 1999 11:32:45 +1000 (EST)

Subject: Re: iron and KWK



1.The goal of feeding centre is to reduce mortality due to global acute malnutrition among children aged 6 months -less than 60 months. Considering that the objective is to reduce the malnutrition prevalence, is there anybody who looked into the possibility that the malnutrition prevalence can be artificially reduced by increase death rate among severely malnourished children or by movement of malnourished population to another area (eg repatriation: case of rwandan refugees in Zaire)

2. I am curious to know if there is a difference in the outcome parametres achieved in a cold climate intervention versus a hot climate intervention for a similar intervention administered? If so what are the explainations?

3. Please tell me if any agency keeps databases on feeding centre management, that is, for malnourished children, admission date, discharge date, diagnosis (Kwash, Marasmus, Kwash-marasmus), Family status (orphans, unaccompanied, both parents alive), referral mode (self-=referred, referred by outreach team, agency etc,), Death and cause of death if any, cause of transfer and other more have been recorded and kept.


Date: Thu, 28 Jan 1999 16:25:01 +0000

From: "Steve Hansch" <>

Subject: RE: camp data



regarding your Question 1:

The research by Phil Nieburg and Angela Berry in Fao camps of Ethiopian refugees in early 1985 found that the apparent rate of malnutrition, measured each month via sample surveys, failed to show a worsening situation (newly malnourished children) because it was masked by an enormously high death rate. The reported malnutrition rate stayed roughly the same over a six month period, even though huge numbers of children were dying; the pool of malnourished being replenished at an equal clip.

Deaths are always comparatively higher among the already-malnourished but in this situation the death rate was of such a high order of magnitude that it significantly distorted the population denominator. In other words, by just looking at periodic sample surveys, we would have thought malnutrition was at least 'under control,' yet in fact it was getting worse during that period.

This masking effect always exists to some degree, though usually small: selective disappearance of a sub-population from observation distorts our summary measures. In a sense it is a sub-class of the phenomenon that epidemiologists refer to as "Competing Risks" where changes in a (e.g. incidence) rate is artificially influenced by shifts in the population denominator of persons at risk for the condition.

I am unaware of any field data that show changes in malnutrition rates due to an observed increase in case-fatality rates.

Disappearance due to repatriation, or even just informal flows into or out of camps, will certainly distort data as well. One unique case was seen in Northern Uganda in 1993: Sudanese refugees "registered" in large reception centers did not stay living in those reception centers; some informally moved around in search of work or back into Sudan. And many of the ill (or most at risk) took advantage of the health facilities in more established refugee villages that maintained registration records separate from the reception centers. As a result, the apparent mortality rate was unusually low in the reception centers and comparatively higher in the village clinics. The mortality rate for the reception center was calculated not based on any true person-year residence in camp but on a wildly inflated figure of how many people were registered there at any time.

question 2:

Except for the IFRC, few humanitarian organizations have systematically studied how standard humanitarian programs differ in cold environments. Of course there are enormous differences, not least of which is the heightened kilocalorie need to compensate for the energy-intensive shiver response.

Food transport on frozen roads (in some cases fuel itself has been known to freeze) pose special concerns not identical to those seen in Africa, Central America or Thailand.

The main outcome parameters are influenced by differing baseline. While Bosnians lost weight during their complex emergency, they had a higher per capita starting weight. Had Ethipians lost as much weight they would have seen deaths from frank starvation, whereas the Bosnians did not see high rates of death due to malnutrition.

question 3:

At the field level, most NGOs do maintain registration data for those malnourished children enrolled in supplementary or therapeutic feeding.

These hand-written records vary from site to site as needs demand and time permits. Rarely are they aggregated into any permanent databases, though sometimes we do review them for long-term lessons (for example, using supplementary feeding program data from CRS, we looked at program efficacy in Somalia in 1993).

In general feeding centers record the data needed to monitor the child's progress toward achieving exit criteria (improvement), which in most situations means comparing their weight/height from week to week, requiring notation of dates, including date of admission. While camp-wide population surveys of malnutrition specifically look for kwashiorkor, it is subject to much inter-observor variation in identification. At the feeding center level, notes will be made of nurse observations, but they are not always against a common/set list of diagnoses. Most of these notes are related to observations directly related to the child's progress, for example, appetite, hookworm infection, signs of corneal scaring, and compliance (showing up each day as expected).

Family data per se is not commonly recorded.

However, separate records may be kept on orphan populations; as in North Kivu (94-96) where distinct centers provided care for "unaccompanied minors" who were more often than not malnourished. During the Ethipian famine, the indigenous Tigreyan relief agency maintained lists and programs for all orphans in the camps.


regards and good luck,

steve hansch

Date: Wed, 27 Jan 1999 10:20:15 +0000

From: George Beaton <>

Subject: climate and treatment


Dear Andre Renzaho,

There is one situation where climate makes a big difference but it is far distant from what Dr. Renzaho has in mind.

In the days when "fortification" of formula with skim milk powder was popular, solute overload became a very serious problem in the summer but not winter. The difference was thought to be water loss by sweat (without solute excretion).

In theory it could happen again but it strikes me that everyone is now very conscious of the issues of solute load and associated water demand.

George Beaton


Note added by Mike Golden:

I saw hypernatraemic dehydration in Mali and Tchad (environmental temp up to 47oC) as I mentioned in a previous message, in children getting a diet with 100kcal/100ml, who where then given additional ReSoMal for mild refeeding diarrhoea. This has not happened since we changed the protocol to 1. make water freely available and restrict oral rehydration fluid to those that needed it (before it was freely available in the ward) and 2. put all children that were judged to require oral rehydration fluid, for whatever reason, back to phase 1 of treatment.

Also there was a real danger of such solute overload with the old TMRU formula which had 135 kcal/100ml and we had to be careful about this formula in any child that had a fever, mild diarrhoea or anorexia and for all small children (less than 4kg) - it was because of the restricted range of patients for whom this diet was suitable and the dangers in places where surveillance and the understanding of the critical nature of the renal solute load could not be assured, that the "standard" diet was changed from the 1981 recommendations to the 100kcal/100 ml of the present "standard" formula. Comparison of the rates of weight gain on the two formula show that they are equivalent. There is no advantage in pushing the energy density of milk-based diets above 100kcal/100ml and it is dangerous.