Calculation of weight gain during catch-up
Calculation of weight gain David Brewster 25.02.2000
Calculation of weight gain Michael Golden 25.02.2000
Weight gain André Briend 25.02.2000


Subject: Treatment of Hypophosphatemia

Date: Fri, 25 Feb 2000 07:21:46 +0930


(Malnourished children often have low serum phosphate levels, and reports by Waterlow, Manary and others have generally found it has prognostic value.

Aboriginal children in northern Australia sometimes have hypophosphatemia, but it seems to be more related to osmotic diarrhoea than severe malnutrition.

My question is at what serum level should treatment be instituted, and what dose of phosphate should be used in children? I cannot get clear indications from the usual sources, so would welcome other's experience and advice.)


An unrelated question: In calculating weight gain in kwashiorkor, I have always subtracted the discharge weight from the lowest weight (after loss of oedema) in grams/ divided by the lowest weight in kg and/ divided again by the length of hospital stay. This is expressed as g/kg/day. Is this other people's understanding. It has been suggested to me that the middle /kg/ should be the discharge weight rather than the lowest weight. Which is correct?



David Brewster

NT Clinical School, Darwin, Australia

Date: Thu, 24 Feb 2000 23:45:25 +0000

From: Michael Golden <>

Subject: Calculation of weight gain.


Dear David,


Assessing weight gain during catch-up.

It is conventional to calculate weight gain in the way that you do it, using the lowest figure and not the discharge or mean Weight. So that you should keep doing it this way for your data to be comparable.

This convention arose for pragmatic reasons - so that rate of weight gain can be calculated before discharge for clinical management of the patient - if they are not gaining 5g/kg/d then there is something wrong - no good to only find this out after discharge.

One problem with this way of measuring weight gain is that the calculation is quite sensitive so changes in the minimum weight and of course the degree of wasting (and hence minimum weight) varies between patients. This is the main reason why people calculate a higher rate of weight gain for marasmus than kwashiorkor - it is a spurious comparison. Also if you measure rate of weight gain over different periods during recovery using different divisors - a steady absolute weight gain would appears to fall off. One way round all these problems is to use the ideal weight as the divisor (100% weight for height) - this then standardizes the reference measures for all the weights gains (and kwashiorkor gains just as quickly as marasmus now - smile) but it does lower the calculated rate of weight gain by at least 30% and people do not like seeing 7g/kg/d instead of 10g/kg/d - the action points etc have been set with the conventional calculation and it is much easier to calculate so I only use the ideal weight divisor when internal comparisons within the data are important - or you want to get a better understanding for what is happening.

Another problem is the "minimum" weight itself - being the lowest weight recorded it is by much more likely to be an erroneous measurement than any of the other measurements taken. ( if you are doing these calculations for scientific comparisons the best way is to regress the subsequent 5 days weights (omitting the minimum weight day - and then take the intercept of the regression at the date of minimum weight - this gets over the problem of a single spurious weight somewhat).

When I use the minimum weigh (or ideal weight) divisor I still get weight-gain falling off as the child recovers. I ascribe this to the child laying down mainly muscle at the beginning and more fat during the later stages of recovery. Actually the mean trajectory follows a power curve - for my data , weight at any time after admission in grams/kg gained was 36.6* days^0.64. It follows from this that the longer period that is considered the slower the rate of weight gain recorded - thus, if one center discharged at 80% weight for height and another at 95% weight for height the first one would have higher rate of weight gain - not because their treatment is any better but because they discharged them before recovery- indeed these patients are more likely to relapse. So when we compare rates of weight gain and use this as a measure of the "success" of a program or particular diet - how we compute that weight gain can have a substantial effect - even if we are all using the "minimum weight"

Mike Golden


Prof. Michael H.N.Golden

Dept of Medicine and Therapeutics, Univ of Aberdeen, Foresterhill, AB9 2ZD. Scotland, (UK)

INTERNET, Tel +44 (1224) 681 818 ext 52793/53014, Fax +44 (1224) 699 884, Tel(direct) +44 (1224) 663 123 527 93

Date: Fri, 25 Feb 2000 11:18:01 +0000

From: "André Briend" <>

Subject: Weight gain


Dear all,

All NGO measure weight gain by a formula assuming that the ratio (weight gain velocity)/weight is constant over time, over a large range of wieghts and >0. Of course, this is not true. If you go back to you math book, you will see this would mean that weight is an exponential function of time. No multicellular organism grow that way. Healthy children do not grow that way either: weight veolicity in g/kg/day is decreaseing all the time with age in pre school children, especially during first year of life, being less than 1 g / kg/ day afer 12 months. No surprise that children do not follow this function for a long period of time. No suprise it does not work when comparing children with different initial nutritional status. No surpise the child weight gain in g/kg/day falls off during recovery, moving to a normal growth pattern, much slower, and always decreasing.

Having said that, this calculation is a good approximation of what happens OVER A SHORT PERIOD OF TIME. Many TFC's in recent years had seen the average weight gain taking off... just after making this simple approximate calculation pointing out some simple feeding problem.

So the message is, CONTINUE TO MEASURE WEIGHT GAIN IN G/KG/DAY, but know it is not an exact description of what happens.

To answer specifically to David question, my remarks apply to whatever weight you take as denominator. The most important is being consistant when making comparisons.

Best regards,



André Briend, MD, PhD CNAM - ISTNA 5 rue du Vertbois, 75003 Paris, France

tel : 33-1-53 01 80 36, fax : 33-1-53 01 80 05