I assume that you are looking for an international reference. I do not think that one exists. The WGS, for example, is for children aged to five years. The previous WHO-endorsed MUAC/A reference was for the same age range. Some countries have elected to create local references based on large cross-sectional surveys. I know of references for Argentina (to 12 years), Bolivia (to 19? years), Turkey (to 17 years), UK (to 18 years), US (to 18 years - this is I think from one of the NHANES rounds), and so on. There may be one for your locale. A search on the WHO site, CDC site, PUBMED or Google should give you an idea of what is available. It would not be very difficult to build your own reference provided age could be ascertained accurately. I think you should be wary of any anthropometric indicator by age since they are very sensitive to even small errors in age and these tend to increase with increasing age. Any anthropometry during adolescence is difficult because maturation does not run to an universal timetable. It is common practice to correct / adjust for sexual maturation rather than age with adolescents (this usually involves examination of breasts and genitals - there are some peculiar measuring devices ... a large measuring spoon for the scrotum is one). I have found these measurements extremely difficult to perform under field conditions. I guess that you could use means to compare populations but, unadjusted for sexual maturation, an individual MUAC would probably have little value. The would be some value in repeated MUAC for treatment monitoring. A suggestion : If anyone knows of a local reference then could they post a link or reference in this thread.

Hi Anonymous 250 There are no age-adjusted MUACS or MUAC z scores for 5-18 year olds. I think they would be immensly valuable for nutrition programmes and HIV clinics. I am currently working on a small project to generate these from NCHS survey data using the same methodology as was used to generate the WHO references. The problem is that, becuase they are a different dataset, they don't join smoothly with the WHO reference values at 5 years - we are working from this - it will probably mean trying some older NCHS data when the US population were a bit thinner. Jay Dr James Berkley MBBS MTropMed MRCP(UK) MD Consultant Paediatrician in Immunology & Infectious Diseases Centre for Geographic Medicine Research (coast) KEMRI/Wellcome Trust PO BOX 230 Kilifi Kenya Tel +254 41 522535 Fax +254 417 522390 Mobile +254 (0) 720 867 011

Jay, You write: "There are no age-adjusted MUACS or MUAC z scores for 5-18 year olds. I think they would be immensly valuable for nutrition programmes and HIV clinics." I wonder how useful given the sexual maturation problem I wrote about. Not such a problem in clinical context particularly with regard to HIV but, I think, a big problem with nutrition programs. I'd appreciate your thoughts.

From Jay: Hi Mark yes, it's a problem. The project did originate from HIV clinic ideas rather than a nutrition programme. But I think it's worht doing it and seeing how it comes out. My (perhaps naiive) thought was that around puberty, the distribution will be broad, which simply makes the z score bands very wide. In fact a premininary look at the data suggest that the distribution is broadest above 20 years rather than in puberty... it may just be a characeristic of the reference population. The only real validation would be against mortality, but that may not be possible in that age group. Once it's worked up a little more, I will email you off-board to discuss. jay

I have just looked through Evelyth and Tanner and there is, perhaps, the makings of an international reference, They report cross-sectional studies of MUAC/A for the required age-group from many locations. If these data could be combined in some sensible way we would have the makings of a "rough and ready" international reference. It is possible that they may report a study local to the original questioner. Jay, It is interesting that these studies show the same pattern as you report.

From Jay: Hi Mark yes, would be good to see how they pan out compared to the NCHS. I;ve just looked at some 37,000 datapoints across all ages from NCHS and it probably does start to spread from around age 15, but the distribution stays very broad after that, i.e. it doesn't narrow again after puberty. jay