1493
Bone turnover in malnourished children
Pyridinoline (PYD) and deoxypyridinoline (DPD) are cross-linking aminoacids of collagen that are located mainly in bone and cartilage. When bone matrix is resorbed these cross-links are quantitatively excreted in the urine and therefore represent specific markers. We have measured the urinary excretion rate of PYD and DPD in 46 severely malnourished boys to assess their skeletal turnover and to relate this to their subsequent rate of growth. The children were aged 13 months (SD 6), and height-for-age was —3·6 (1·6) Z-score, and weight-for-height was —2·4 (0·8) Z-score. PYD excretion when malnourished and after "recovery" was 11·2 (4·6) nmol h-1m-2 and 32·2 (10·8) nmol h-1m-2 and DPD excretion was 2·6 (1·3) nmol h-1m-2 and 7·5 (3·0) nmol h-1m-2, respectively. The ratio of the two cross-links did not change with recovery. These data show that cartilage and bone turnover is much lower in the malnourished than in the recovered child. There was no difference in the degree of depression of turnover between the children with marasmus, marasmic-kwashiorkor, or kwashiorkor. The rate of height gain during recovery was significantly related to cross-link excretion, age, and weight-for-height on admission. These three factors accounted for 44% of the variance in the height velocity of the children. PYD and DPD excretion rate could be used to assess therapeutic interventions designed to alleviate
cartilage (PYD).’ PYD and DPD released from bone during resorption are not metabolised and dietary-derived PYD and DPD are not absorbed. Although small amounts of cross-links are present in other tissues, such as tendons, ligaments, and the aorta, these tissues have a low turnover so their contribution to the urinary pool is small. Thus, excretion of PYD and DPD is directly related to the amount of bone resorbed and, hence, to bone turnover.g This has been confirmed in adults by bone histomorphometry9 and by the close relation between cross-link output and bone turnover measured isotopically.10 We have established a specific assay" for these cross-links and have used this method to determine possible changes in bone resorption during malnutrition and whether the resulting rate of bone and
turnover
is related to recovery.
Subjects and methods 46 malnourished boys, aged 13 (SD 6) months, were investigated the Tropical Metabolism Research Unit, University of the West Indies, Jamaica. Table I shows diagnoses, according to the Wellcome classification,12 and anthropometric characteristics. The children with marasmus or marasmic-kwashiorkor were more stunted than those with kwashiorkor. They were treated by standard methods13 including an energy-dense diet based on cow’s milk formulated from Perlargon (Nestle) and corn or coconut oil. at
TABLE I-PATIENT DETAILS AT ADMISSION AND DISCHARGE, ACCORDING TO TYPE OF MALNUTRITION
stunting. Lancet 1992; 340: 1493-96.
Introduction Of the various forms of malnutrition, stunting (nutritional dwarfing) is by far the most prevalent; about half the world’s children are affected. In most poor countries and in disadvantaged sections of affluent societies they grow into adults of reduced stature. The relative importance of various insults and the pathological mechanisms of stunting are not known. Radiologically the bones of malnourished children are short, but their width seems to be in proportion to their length. Reduced bone density and thickness is nearly always reported. Whether this is due to an increase in bone resorption with the onset of malnutrition or to decreased synthesis is controversial.2,3 The excretion of hydroxyproline is reduced in severe forms of malnutrition,4,s but this substance is not a reliable marker of bone collagen turnover in this disease because it is present in all types of collagen, much of which is in soft tissues. The interpretation of hydroxyproline excretion in relation to body weight is further complicated by the changes in body composition ; in soft tissues, the proportion of collagen to non-collagen protein increases substantially in malnutrition.6 Pyridinoline (PYD) and deoxypyridinoline (DPD) are cross-linking aminoacids that are formed during the maturation of collagen fibrils in bone (PYD and DPD)
Anthropometric variables were calculated with National Center for Health Statistics standards and expressed m standard deviation scores (Z)." *Differences among the three malnourished groups tTime from minimum to maximum weight in ward. tTlme from admission to start of accelerated weight gain Mar-kwash = marasmic kwashiokor, NS = not significant Means (SD) shown
The children took an average of 4 days to be resuscitated and to start their accelerated weight gain; the period of active growth was 36 (13) days by which time they were at the reference weight for a normal ADDRESSES: Istituto Nazionale della Nutrizione, Rome, Italy (F Branca, MD, Prof A. Ferro-Luzzi, MD); Rowett Research Institute, Aberdeen, UK (S. P. Robins, DSc, Prof M H. N Golden, FRCP); and Tropical Metabolism Research Unit, University of the West Indies, Jamaica (Prof M. H. N Golden) Correspondence to Prof M. H N Golden, Department of Medicine and Therapeutics, University of Aberdeen, Foresterhill, Aberdeen AB9 2ZD, UK.
1494
TABLE II-MULTIPLE REGRESSION ANALYSIS OF RATE OF HEIGHT GAIN BY CHILDREN RECOVERING FROM MALNUTRITION
RHG =0-026 (0.129) -0.021 (0 005) age + 0 020 (0-006) PYD - 0-116 (0 037)WFH. R HG == rate of height gain, WFH = weight-for-height, where age is in months, PYD in nmol h-1m-2, and WFH in Z score unrts. Error values are SE of the estimates of the coefficients. There was no relation between rates of height and weight gain, PYD excretion at discharge, or height-for-age (degree of stuntmg) Vanance explained R2 x 100. =
expressed per unit skeletal mass. To determine the most appropriate surrogate for skeletal mass, logarithmetric regressions of absolute excretion of cross-links against height were done. The exponents for PYD were 18 (SE of estimate 0-6) at admission and 22 (05) at discharge; the corresponding exponents derived from DPD were 1-9 (0-7) and 2-5 (0-5). Because none of the exponents were significantly different from 2, the square of a patient’s height was used as a surrogate for skeletal mass in this population. The cross-link excretion data were expressed per unit height2 for all subsequent analyses.
Pyridinoline rates in
and deoxypyridinoline
(mmol
h-1
m-2) excretion
malnourished children on admission and at discharge.
Box plots show median (horizontal line), interquartile range (shaded), and range (bars). K=kwashiorkor; M=marasmus; MK=marasmickwashiorkor.
child of the same height (table i). The children were in hospital for a total of 49 (18) days. Accurately timed urine collections, of about 24 h, were started shortly after admission and again when the child had reached the weight appropriate for height. Urine was collected by continuous aspiration of a plastic bag attached to the child’s perineum by silicone-based colostomy cement, through 3 m lengths of Tygon tubing. IS The children were not restrained during urine collection, but were confined to a cot. Urine was aspirated into a cooled collecting vessel containing solid chlorhexidine, and samples were stored at -20°C until analysis. PYD and DPD were assayed by reverse-phase high-pressure liquid chromatography on urine hydrolysates8,11 with heptafluorobutyric acid as the ion pairing agent. Samples were tested in duplicate, with a mean coefficient of variation of 4%. Previous experiments8 showed that storage at —20°C does not affect the assay. Anthropometric measurements were done by nursing staff with standard procedures. The length board was calibrated to 0 1 cm and electronic scales to 1 g. Statistical analysis was done with Systat.16 Analysis of variance and multiple regression analysis were used where appropriate. Statistical significance was accepted when p < 0 05. ’
Results The absolute excretion rates of both PYD and DPD were linearly related to the measures of the children’s sizeweight, height, and functions of height-at both admission and at discharge. The excretion data should ideally be
When children were malnourished, the rate of excretion of PYD and DPD was about a third of that after recovery (figure). Cross-link output did not differ (p< 0-001) between the diagnostic categories at admission or at discharge. Within the group as a whole, the rate of cross-link excretion was not related to the children’s age, weight, degree of wasting (weight-for-height), or deficit in height (height-for-age). The deficit in height is presumably related to the time that the child had growth arrest. The degree of depression of bone turnover in the malnourished patients might have been related to severity of illness other than wasting or stunting. As a proxy for degree of illness we used both the time taken to overcome anorexia and start gaining weight and the time for the patient to recover normal body proportions. Neither of these measures of illness was related to amount of cross-link excretion. The predominance of PYD in cartilage in contrast with the presence of both cross-links in bone 17 suggests that any alteration in the relative turnover of bone and cartilage would result in changes in the relative proportions of the cross-links excreted. There were, however, no significant changes in the PYD/DPD ratio between admission values of 4-6 (SD 1-2), 4-2 (0-7), and 4-6 (1-1) for children with marasmus, marasmickwashiorkor, and kwashiorkor, respectively, and discharge values of 4-4 (0-5), 4-5 (0-7), and 4-1 (0-8), respectively, indicating that no differential changes in endochondral growth relative to bone remodelling had occurred. The variability among individual children in the rate of cross-link excretion was much higher after recovery than when they were malnourished (figure). At recovery, the rate of excretion of cross-links was not related to previous diagnosis, age, weight-for-height, height-for-age, or time taken to recover. During treatment, the children gained about a third of their admission body weight (table I) to achieve normal body proportions (weight-for-height). There was no relation between the intensity of weight gain and bone turnover. The children did not reduce their deficit in height-for-age (table I). The rate of height gain during recovery was positively related to age, bone turnover at time of admission, and degree of wasting (weight-for-height); altogether, 44% )fthe variability in the height response of the children to the hospital phase of rehabilitation could be ascribed to these hree factors (table II).
1495
TABLE III-BONE RESORPTION RATES IN MALNOURISHED AND RECOVERED CHILDREN
Values (mean, SE of and that the skeletal
estimates) mass is
assume
that 1 nmol D PD
corresponds to 17 mg bone
200 g
On the assumption that the DPD content of bone in our children was 0-07 mol per mol collagen,18 and that collagen represents 25% by weight of hydrated mineralised bone, we estimate that excretion of 1 nmol DPD corresponds to 17 mg of bone resorbed. 19 Table III shows the calculated amount of bone resorbed in the children.
Discussion Unlike other aspects of malnutrition the dynamics of stunting and subsequent catch-up have received little attention. The control of growth with respect to interrelations of nutritional intake, infection, and hormonal balance is not understood. The major impediment has been the difficulty in measuring changes in bone on the same time scale as that for nutrient intake and balance, infectious activity, or amount of cytokine and hormone secretion. Although impressive gains in height have been recorded when short children are given milk supplements or treatment for longstanding infections such as trichuris dysentry syndrome,21 observations such as these have been made over 6 months or more.22 The use of PYD and DPD as specific and quantitative measures of rate of bone and cartilage breakdown has enabled us to show that the turnover of the skeleton is greatly reduced in malnutrition and that it responds rapidly to nutritional intervention. The deficit in height of the children is probably a measure of the duration of their impaired growth. Since the severity of stunting was not related to the reduction in bone turnover, the metabolic abnormality in bone does not seem to become progressively more severe. Further, because the increase in turnover with treatment was substantial in both the more and the less stunted children, the longstanding growth retardation was not due to an inherent defect in the child, nor did it lead to permanent reduction in bone metabolic
activity. We used the children as their own controls, but were the values obtained after recovery the same as those in children who had never been malnourished? We have two reasons to suppose that they are. First, in a group of 103 Italian children aged 3-5 years PYD output was 65% (21 2 nmol h-lm-2 [SD 8.1] vs 32-2 [10-8]), and DPD was 90% (6-7 nmolh-lm-2 [2’9] vs 75 [3’0]) of that found in our groups of recovered children (unpublished results): this is the magnitude of the difference that we would anticipate from the change in cross-link output with age, expressed per unit creatinine (unpublished work). Second, the previously malnourished children were not growing at rates above those expected for their age or height.23 Indeed, Prudhon et al24 demonstrated very low serum osteocalcin (a marker of bone formation) values in children with kwashiorkor in Senegal which increased to about half the control value after 3 weeks’ treatment; there was no indication that osteocalcin output increased to supranormal values in these children. The total weight of the skeleton in a healthy 12-monthold child is about 200 g,25 and from our calculations of bone resorption rates in malnourished and recovered children
(table III), we deduce that the malnourished child turns over %, and the recovered child 0-7%, of bone mass per
about 0-2
day. Although in this study the children gained height at about the same rate as did those in our previous study,26 it should be emphasised that much of the variance in the rate of height gain is probably caused by errors in measurement of changes in height over such short periods. In view of the measurement errors, the explanation of 44% of the variance in height gain on the basis of three measured variables probably underestimates their actual contribution. Age and weight-for-height were also related to height gain in our previous study.26 The younger children grew more rapidly than older children. The degree of wasting is related to the height gain because the children seem to make up most of their weight deficits before starting to grow in height. This discordance between weight and height gain has been seen in community supplementation trials22 and in the seasonality of wasting and Stunting.27 In addition, the degree of depression of bone turnover, which accompanies malnutrition, is independently related to the rate of gain in height during the ensuing period of rehabilitation. That the level of bone turnover is higher in children who gain most height is analogous to the relation between rate of turnover of body protein and of gain in weight.28 The demonstration that PYD and DPD are related to the of longitudinal growth in children recovering from malnutrition should lead to studies that examine the effect of dietary manipulation on longitudinal growth so that the requirements for longitudinal growth can be defined and steps taken to prevent growth retardation. rate
F. B. was supported by a grant from the British Council, and M. H. N. G. by the Wellcome Trust. We thank Dr J. Doherty, Dr D. Ramdath, and the nursing staff of the Tropical Metabolism Research Unit, University of the West Indies, Jamaica, for help with the urine collections. We also acknowledge support from the Scottish Office Agriculture and Fisheries Department.
REFERENCES
JH. Bone growth and development in protein-calorie malnutrition. Wold Rev Nutr Diet 1978; 28: 143-87. 2. Garn SM. Contributions of the radiographic image to our knowledge of 1. Himes
3.
human growth. Am J Roentgenol 1981; 137: 231-39. Akamaguna AI, Odita JC, Ugbodaga CI, Okolo AA. Bone and soft tissue components of the leg in infants with protein calorie malnutrition.
Pediatr Radiol 1986; 16: 40-42. 4. Picou DIM, Alleyne GAO, Seakins A. Hydroxyproline and creatinine excretion in infantile protein malnutrition. Clin Sci 1965; 29: 517-23. 5. Whitehead RG. Hydroxyproline creatinine ratio as an index of nutritional status and rate of growth. Lancet 1965; ii: 567-70. 6. Picou DIM, Halliday D, Garrow JS. Total body protein, collagen and non-collagen protein in infantile protein malnutrition. Clin Sci 1966; 30: 345-51. 7. Robins SP. Functional properties of collagen and elastin. Baillières Clin Rheumatol 1988; 2: 1-36. 8. Robins SP, Black D, Paterson CR, Reid DM, Duncan A, Seibel MJ. Evaluation of urinary hydroxypyridinium crosslink measurements as resorption markers in metabolic bone diseases. Eur J Clin Invest 1991; 21: 310-15. 9. Delmas PR, Schlemmer A, Gineyts E, Riis B, Christiansen C. Urinary excretion of pyridinoline crosslinks correlates with bone turnover measured on iliac crest biopsy in patients with vertebral osteoporosis. J Bone Mineral Res 1991; 6: 639-44. 10. Eastell R, Hampton L, Colwell A, et al. Urinary collagen crosslinks are highly correlated with radioisotopic measurements of bone resorption. In: Christiansen C, Overgeard K, eds. Proceedings of the Third International Symposium on Osteoporosis, Copenhagen, 1990. Copenhagen: Osteopress ApS, 1990: 469-70. 11. Black D, Duncan A, Robins SP. Quantitative analysis of the pyridinium crosslinks of collagen in urine using ion-paired reversed-phase high-performance liquid chromatography. Anal Biochem 1988; 169: 197-203. 12. Anon. Classification of infantile malnutrition Lancet 1970; ii: 302-03.
1496
13.
Jackson AA, Golden MHN. Severe malnutrition.In: Weatherall DJ, Ledington JGG, Warrell DA, eds. The Oxford textbook of medicine, 2nd ed. Oxford: Oxford University Press, 1987: 8.12-8.23. 14. Hamill PVV, Dridz TA, Johnson CZ, Reed RB, Roche AF, Moore WM. Physical growth: National Center for Health Statistics percentiles. Am J Clin Nutr 1979; 32: 607-29. 15. Golden MHN, Golden BE, Jackson AA. Method of trace metal balance studies in young children. In: Howell McC J, Gawthorne JM, White CH, eds. Trace element metabolism in animals and man 4. Canberra: Australian Academy of Science, 1981: 69-72. 16. Wilkinson L. Systat: the system for statistics. Evanston, Illinois: Systat, 1990. 17.
Eyre DR, Koob TJ, Van Ness KP. Quantification of hydroxypyridinium crosslinks in collagen by high-performance liquid chromatography.
Anal Biochem 1984; 137: 380-88. Eyre DR, Dickson IR, Van Ness KP. Collagen crosslinking in human bone and cartilage: age related changes in the content of mature hydroxypyridinium residues. Biochem J 1988; 252: 495-500. 19. Beardsworth LJ, Eyre DR, Dickson IR. Changes with age in the urinary excretion of lysyl- and hydroxylysylpyridinoline, two new markers of bone collagen turnover. J Bone Mineral Res 1990; 5: 671-76. 20. Lampl M, Johnston FE, Malcolm LA. The effects of protein supplementation on the growth and skeletal maturation of New 18.
Guinean school children. Ann Hum Biol 1978; 5: 219-27.
21.
Cooper ES, Bundy DAP, MacDonald TT, Golden MHN. Growth suppression in the trichuris dysentry syndrome. Eur J Clin Nutr 1990;
44: 285-91. 22. Heikens GT, Schofield WN, Dawson S, Grantham-McGregor SM. The Kingston project I, growth of malnourished children during rehabilitation in the community, given a high energy supplement. Eur J Clin Nutr 1989; 43: 145-60. 23. Baumgartner RN, Roche AF, Himes JH. Incremental growth tables: supplementary to previously published charts. Am J Clin Nutr 1986; 43: 711-22. 24. Prudhon C, Sall G, Ndiaye B, Lemonnier D. Nutritional regulation of serum osteocalcin: study in kwashiorkor. C R Acad Sci III 1991; 313: 233-38. 25. Trotter M, Peterson RR. Weight of the skeleton during postnatal development. Am J Phys Anthropol 1970; 33: 313-24. 26. Walker SP, Golden MHN. Growth in length of children recovering from severe malnutrition. Eur J Clin Nutr 1988; 42: 395-404.
KH, Black RE, Becker S. Seasonal changes in nutritional status and the prevalence of malnutrition in a longitudinal study of young children in rural Bangladesh. Am J Clin Nutr 1982; 36: 303-13. 28. Golden MHN, Waterlow JC, Picou DIM. The relationship between dietary intake, weight change, nitrogen balance and protein turnover in man. Am J Clin Nutr 1977; 30: 1345-48. 27. Brown
Comparison of saliva and serum for HIV surveillance in developing countries
Saliva has been proposed as a non-invasive alternative to serum for HIV antibody testing. In a field study in Myanmar (formerly Burma), we evaluated such an alternative to identify the frequency of HIV infection in a surveillance programme of high-risk and low-risk sentinel groups. Duplicate vials of saliva and serum were collected from 479 high-risk and 1039 low-risk subjects. One vial of each pair was analysed blind in two laboratories, one in the USA and the other in Myanmar. The US laboratory followed WHO confirmatory strategy II I with three different enzymelinked immunosorbent assays (ELISAs), while the laboratory in Myanmar followed strategy I with one ELISA. Serum testing in the US was the gold standard. The Cambridge ELISA with saliva was a more effective surveillance tool (sensitivity 90·5%, specificity 99·5-100%) for describing the frequency of subjects with HIV antibodies than the serum ELISA supplied to Myanmar by WHO (95·9% and 98·3%, respectively). Saliva is recommended as a safe and effective alternative to serum for HIV antibody testing with ELISA in surveillance programmes in developing countries. ..-
.
v
Lancet 1992; 340: 1496-99.
Introduction Infection with HIV-1 is
becoming evident in Myanmar The (formerly Burma). frequency of HIV-1 infection has risen dramatically in Myanmar to 2-3% during 1990-92. Starting in April-May, 1992, a sentinel surveillance
programme was established, modelled after the successful programme in Thailand.l About 100 subjects are selected every 6 months from high-risk and low-risk groups at various sites. In addition, information is gathered on all blood donors at selected sites. The World Health Organisation (WHO) provides most of the HIV kits but because the number of kits is limited, the sentinel sites and risk groups have to be carefully selected to provide maximum information. The first round of the programme showed that the HIV epidemic is spreading in intravenous drug users, has infected sexually active individuals in middle-risk groups (patients at sexually transmitted disease [STD] clinics, prostitutes) and a low-risk group (women attending antenatal clinics), and is spreading to the general
community (blood donors) (fig 1). If the epidemic in Myanmar is to be controlled, testing needs to be more widespread to make people aware that HIV has reached their community. Yet for Myanmar, as for many developing countries, the cost of widespread testing is too great. Test kits require foreign currency, and personnel to draw blood need to be trained from people who may fear HIV infection from needles. Saliva has been suggested as a non-invasive alternative for HIV testing2-6 in surveillance programmes and in community-based surveys, although in a developing country, climate and conditions are far from ideal for saliva tests.5,6 To
assess
the value of saliva for
surveillance, we did a field study in Myanmar. ADDRESSES: Department of Epidemiology, University of California at Los Angeles (Prof R. R. Frerichs, DrPH); AIDS Prevention and Control Programme, Department of Health, Yangon, Myanmar (M. T. Htoon, MBBS); Saliva Diagnostic Systems, Vancouver, Washington, USA (N. Eskes, BS); and National Health Laboratory, Yangon, Myanmar (S. Lwin, MBBS). Correspondence to Prof R. R. Frerichs, Department of Epidemiology, UCLA School of Public Health, Los Angeles, California 90024-1772, USA.
Bone turnover in malnourished children
Pyridinoline (PYD) and deoxypyridinoline (DPD) are cross-linking aminoacids of collagen that are located mainly in bone and cartilage. When bone matrix is resorbed these cross-links are quantitatively excreted in the urine and therefore represent specific markers. We have measured the urinary excretion rate of PYD and DPD in 46 severely malnourished boys to assess their skeletal turnover and to relate this to their subsequent rate of growth. The children were aged 13 months (SD 6), and height-for-age was —3·6 (1·6) Z-score, and weight-for-height was —2·4 (0·8) Z-score. PYD excretion when malnourished and after "recovery" was 11·2 (4·6) nmol h-1m-2 and 32·2 (10·8) nmol h-1m-2 and DPD excretion was 2·6 (1·3) nmol h-1m-2 and 7·5 (3·0) nmol h-1m-2, respectively. The ratio of the two cross-links did not change with recovery. These data show that cartilage and bone turnover is much lower in the malnourished than in the recovered child. There was no difference in the degree of depression of turnover between the children with marasmus, marasmic-kwashiorkor, or kwashiorkor. The rate of height gain during recovery was significantly related to cross-link excretion, age, and weight-for-height on admission. These three factors accounted for 44% of the variance in the height velocity of the children. PYD and DPD excretion rate could be used to assess therapeutic interventions designed to alleviate
cartilage (PYD).’ PYD and DPD released from bone during resorption are not metabolised and dietary-derived PYD and DPD are not absorbed. Although small amounts of cross-links are present in other tissues, such as tendons, ligaments, and the aorta, these tissues have a low turnover so their contribution to the urinary pool is small. Thus, excretion of PYD and DPD is directly related to the amount of bone resorbed and, hence, to bone turnover.g This has been confirmed in adults by bone histomorphometry9 and by the close relation between cross-link output and bone turnover measured isotopically.10 We have established a specific assay" for these cross-links and have used this method to determine possible changes in bone resorption during malnutrition and whether the resulting rate of bone and
turnover
is related to recovery.
Subjects and methods 46 malnourished boys, aged 13 (SD 6) months, were investigated the Tropical Metabolism Research Unit, University of the West Indies, Jamaica. Table I shows diagnoses, according to the Wellcome classification,12 and anthropometric characteristics. The children with marasmus or marasmic-kwashiorkor were more stunted than those with kwashiorkor. They were treated by standard methods13 including an energy-dense diet based on cow’s milk formulated from Perlargon (Nestle) and corn or coconut oil. at
TABLE I-PATIENT DETAILS AT ADMISSION AND DISCHARGE, ACCORDING TO TYPE OF MALNUTRITION
stunting. Lancet 1992; 340: 1493-96.
Introduction Of the various forms of malnutrition, stunting (nutritional dwarfing) is by far the most prevalent; about half the world’s children are affected. In most poor countries and in disadvantaged sections of affluent societies they grow into adults of reduced stature. The relative importance of various insults and the pathological mechanisms of stunting are not known. Radiologically the bones of malnourished children are short, but their width seems to be in proportion to their length. Reduced bone density and thickness is nearly always reported. Whether this is due to an increase in bone resorption with the onset of malnutrition or to decreased synthesis is controversial.2,3 The excretion of hydroxyproline is reduced in severe forms of malnutrition,4,s but this substance is not a reliable marker of bone collagen turnover in this disease because it is present in all types of collagen, much of which is in soft tissues. The interpretation of hydroxyproline excretion in relation to body weight is further complicated by the changes in body composition ; in soft tissues, the proportion of collagen to non-collagen protein increases substantially in malnutrition.6 Pyridinoline (PYD) and deoxypyridinoline (DPD) are cross-linking aminoacids that are formed during the maturation of collagen fibrils in bone (PYD and DPD)
Anthropometric variables were calculated with National Center for Health Statistics standards and expressed m standard deviation scores (Z)." *Differences among the three malnourished groups tTime from minimum to maximum weight in ward. tTlme from admission to start of accelerated weight gain Mar-kwash = marasmic kwashiokor, NS = not significant Means (SD) shown
The children took an average of 4 days to be resuscitated and to start their accelerated weight gain; the period of active growth was 36 (13) days by which time they were at the reference weight for a normal ADDRESSES: Istituto Nazionale della Nutrizione, Rome, Italy (F Branca, MD, Prof A. Ferro-Luzzi, MD); Rowett Research Institute, Aberdeen, UK (S. P. Robins, DSc, Prof M H. N Golden, FRCP); and Tropical Metabolism Research Unit, University of the West Indies, Jamaica (Prof M. H. N Golden) Correspondence to Prof M. H N Golden, Department of Medicine and Therapeutics, University of Aberdeen, Foresterhill, Aberdeen AB9 2ZD, UK.
1494
TABLE II-MULTIPLE REGRESSION ANALYSIS OF RATE OF HEIGHT GAIN BY CHILDREN RECOVERING FROM MALNUTRITION
RHG =0-026 (0.129) -0.021 (0 005) age + 0 020 (0-006) PYD - 0-116 (0 037)WFH. R HG == rate of height gain, WFH = weight-for-height, where age is in months, PYD in nmol h-1m-2, and WFH in Z score unrts. Error values are SE of the estimates of the coefficients. There was no relation between rates of height and weight gain, PYD excretion at discharge, or height-for-age (degree of stuntmg) Vanance explained R2 x 100. =
expressed per unit skeletal mass. To determine the most appropriate surrogate for skeletal mass, logarithmetric regressions of absolute excretion of cross-links against height were done. The exponents for PYD were 18 (SE of estimate 0-6) at admission and 22 (05) at discharge; the corresponding exponents derived from DPD were 1-9 (0-7) and 2-5 (0-5). Because none of the exponents were significantly different from 2, the square of a patient’s height was used as a surrogate for skeletal mass in this population. The cross-link excretion data were expressed per unit height2 for all subsequent analyses.
Pyridinoline rates in
and deoxypyridinoline
(mmol
h-1
m-2) excretion
malnourished children on admission and at discharge.
Box plots show median (horizontal line), interquartile range (shaded), and range (bars). K=kwashiorkor; M=marasmus; MK=marasmickwashiorkor.
child of the same height (table i). The children were in hospital for a total of 49 (18) days. Accurately timed urine collections, of about 24 h, were started shortly after admission and again when the child had reached the weight appropriate for height. Urine was collected by continuous aspiration of a plastic bag attached to the child’s perineum by silicone-based colostomy cement, through 3 m lengths of Tygon tubing. IS The children were not restrained during urine collection, but were confined to a cot. Urine was aspirated into a cooled collecting vessel containing solid chlorhexidine, and samples were stored at -20°C until analysis. PYD and DPD were assayed by reverse-phase high-pressure liquid chromatography on urine hydrolysates8,11 with heptafluorobutyric acid as the ion pairing agent. Samples were tested in duplicate, with a mean coefficient of variation of 4%. Previous experiments8 showed that storage at —20°C does not affect the assay. Anthropometric measurements were done by nursing staff with standard procedures. The length board was calibrated to 0 1 cm and electronic scales to 1 g. Statistical analysis was done with Systat.16 Analysis of variance and multiple regression analysis were used where appropriate. Statistical significance was accepted when p < 0 05. ’
Results The absolute excretion rates of both PYD and DPD were linearly related to the measures of the children’s sizeweight, height, and functions of height-at both admission and at discharge. The excretion data should ideally be
When children were malnourished, the rate of excretion of PYD and DPD was about a third of that after recovery (figure). Cross-link output did not differ (p< 0-001) between the diagnostic categories at admission or at discharge. Within the group as a whole, the rate of cross-link excretion was not related to the children’s age, weight, degree of wasting (weight-for-height), or deficit in height (height-for-age). The deficit in height is presumably related to the time that the child had growth arrest. The degree of depression of bone turnover in the malnourished patients might have been related to severity of illness other than wasting or stunting. As a proxy for degree of illness we used both the time taken to overcome anorexia and start gaining weight and the time for the patient to recover normal body proportions. Neither of these measures of illness was related to amount of cross-link excretion. The predominance of PYD in cartilage in contrast with the presence of both cross-links in bone 17 suggests that any alteration in the relative turnover of bone and cartilage would result in changes in the relative proportions of the cross-links excreted. There were, however, no significant changes in the PYD/DPD ratio between admission values of 4-6 (SD 1-2), 4-2 (0-7), and 4-6 (1-1) for children with marasmus, marasmickwashiorkor, and kwashiorkor, respectively, and discharge values of 4-4 (0-5), 4-5 (0-7), and 4-1 (0-8), respectively, indicating that no differential changes in endochondral growth relative to bone remodelling had occurred. The variability among individual children in the rate of cross-link excretion was much higher after recovery than when they were malnourished (figure). At recovery, the rate of excretion of cross-links was not related to previous diagnosis, age, weight-for-height, height-for-age, or time taken to recover. During treatment, the children gained about a third of their admission body weight (table I) to achieve normal body proportions (weight-for-height). There was no relation between the intensity of weight gain and bone turnover. The children did not reduce their deficit in height-for-age (table I). The rate of height gain during recovery was positively related to age, bone turnover at time of admission, and degree of wasting (weight-for-height); altogether, 44% )fthe variability in the height response of the children to the hospital phase of rehabilitation could be ascribed to these hree factors (table II).
1495
TABLE III-BONE RESORPTION RATES IN MALNOURISHED AND RECOVERED CHILDREN
Values (mean, SE of and that the skeletal
estimates) mass is
assume
that 1 nmol D PD
corresponds to 17 mg bone
200 g
On the assumption that the DPD content of bone in our children was 0-07 mol per mol collagen,18 and that collagen represents 25% by weight of hydrated mineralised bone, we estimate that excretion of 1 nmol DPD corresponds to 17 mg of bone resorbed. 19 Table III shows the calculated amount of bone resorbed in the children.
Discussion Unlike other aspects of malnutrition the dynamics of stunting and subsequent catch-up have received little attention. The control of growth with respect to interrelations of nutritional intake, infection, and hormonal balance is not understood. The major impediment has been the difficulty in measuring changes in bone on the same time scale as that for nutrient intake and balance, infectious activity, or amount of cytokine and hormone secretion. Although impressive gains in height have been recorded when short children are given milk supplements or treatment for longstanding infections such as trichuris dysentry syndrome,21 observations such as these have been made over 6 months or more.22 The use of PYD and DPD as specific and quantitative measures of rate of bone and cartilage breakdown has enabled us to show that the turnover of the skeleton is greatly reduced in malnutrition and that it responds rapidly to nutritional intervention. The deficit in height of the children is probably a measure of the duration of their impaired growth. Since the severity of stunting was not related to the reduction in bone turnover, the metabolic abnormality in bone does not seem to become progressively more severe. Further, because the increase in turnover with treatment was substantial in both the more and the less stunted children, the longstanding growth retardation was not due to an inherent defect in the child, nor did it lead to permanent reduction in bone metabolic
activity. We used the children as their own controls, but were the values obtained after recovery the same as those in children who had never been malnourished? We have two reasons to suppose that they are. First, in a group of 103 Italian children aged 3-5 years PYD output was 65% (21 2 nmol h-lm-2 [SD 8.1] vs 32-2 [10-8]), and DPD was 90% (6-7 nmolh-lm-2 [2’9] vs 75 [3’0]) of that found in our groups of recovered children (unpublished results): this is the magnitude of the difference that we would anticipate from the change in cross-link output with age, expressed per unit creatinine (unpublished work). Second, the previously malnourished children were not growing at rates above those expected for their age or height.23 Indeed, Prudhon et al24 demonstrated very low serum osteocalcin (a marker of bone formation) values in children with kwashiorkor in Senegal which increased to about half the control value after 3 weeks’ treatment; there was no indication that osteocalcin output increased to supranormal values in these children. The total weight of the skeleton in a healthy 12-monthold child is about 200 g,25 and from our calculations of bone resorption rates in malnourished and recovered children
(table III), we deduce that the malnourished child turns over %, and the recovered child 0-7%, of bone mass per
about 0-2
day. Although in this study the children gained height at about the same rate as did those in our previous study,26 it should be emphasised that much of the variance in the rate of height gain is probably caused by errors in measurement of changes in height over such short periods. In view of the measurement errors, the explanation of 44% of the variance in height gain on the basis of three measured variables probably underestimates their actual contribution. Age and weight-for-height were also related to height gain in our previous study.26 The younger children grew more rapidly than older children. The degree of wasting is related to the height gain because the children seem to make up most of their weight deficits before starting to grow in height. This discordance between weight and height gain has been seen in community supplementation trials22 and in the seasonality of wasting and Stunting.27 In addition, the degree of depression of bone turnover, which accompanies malnutrition, is independently related to the rate of gain in height during the ensuing period of rehabilitation. That the level of bone turnover is higher in children who gain most height is analogous to the relation between rate of turnover of body protein and of gain in weight.28 The demonstration that PYD and DPD are related to the of longitudinal growth in children recovering from malnutrition should lead to studies that examine the effect of dietary manipulation on longitudinal growth so that the requirements for longitudinal growth can be defined and steps taken to prevent growth retardation. rate
F. B. was supported by a grant from the British Council, and M. H. N. G. by the Wellcome Trust. We thank Dr J. Doherty, Dr D. Ramdath, and the nursing staff of the Tropical Metabolism Research Unit, University of the West Indies, Jamaica, for help with the urine collections. We also acknowledge support from the Scottish Office Agriculture and Fisheries Department.
REFERENCES
JH. Bone growth and development in protein-calorie malnutrition. Wold Rev Nutr Diet 1978; 28: 143-87. 2. Garn SM. Contributions of the radiographic image to our knowledge of 1. Himes
3.
human growth. Am J Roentgenol 1981; 137: 231-39. Akamaguna AI, Odita JC, Ugbodaga CI, Okolo AA. Bone and soft tissue components of the leg in infants with protein calorie malnutrition.
Pediatr Radiol 1986; 16: 40-42. 4. Picou DIM, Alleyne GAO, Seakins A. Hydroxyproline and creatinine excretion in infantile protein malnutrition. Clin Sci 1965; 29: 517-23. 5. Whitehead RG. Hydroxyproline creatinine ratio as an index of nutritional status and rate of growth. Lancet 1965; ii: 567-70. 6. Picou DIM, Halliday D, Garrow JS. Total body protein, collagen and non-collagen protein in infantile protein malnutrition. Clin Sci 1966; 30: 345-51. 7. Robins SP. Functional properties of collagen and elastin. Baillières Clin Rheumatol 1988; 2: 1-36. 8. Robins SP, Black D, Paterson CR, Reid DM, Duncan A, Seibel MJ. Evaluation of urinary hydroxypyridinium crosslink measurements as resorption markers in metabolic bone diseases. Eur J Clin Invest 1991; 21: 310-15. 9. Delmas PR, Schlemmer A, Gineyts E, Riis B, Christiansen C. Urinary excretion of pyridinoline crosslinks correlates with bone turnover measured on iliac crest biopsy in patients with vertebral osteoporosis. J Bone Mineral Res 1991; 6: 639-44. 10. Eastell R, Hampton L, Colwell A, et al. Urinary collagen crosslinks are highly correlated with radioisotopic measurements of bone resorption. In: Christiansen C, Overgeard K, eds. Proceedings of the Third International Symposium on Osteoporosis, Copenhagen, 1990. Copenhagen: Osteopress ApS, 1990: 469-70. 11. Black D, Duncan A, Robins SP. Quantitative analysis of the pyridinium crosslinks of collagen in urine using ion-paired reversed-phase high-performance liquid chromatography. Anal Biochem 1988; 169: 197-203. 12. Anon. Classification of infantile malnutrition Lancet 1970; ii: 302-03.
1496
13.
Jackson AA, Golden MHN. Severe malnutrition.In: Weatherall DJ, Ledington JGG, Warrell DA, eds. The Oxford textbook of medicine, 2nd ed. Oxford: Oxford University Press, 1987: 8.12-8.23. 14. Hamill PVV, Dridz TA, Johnson CZ, Reed RB, Roche AF, Moore WM. Physical growth: National Center for Health Statistics percentiles. Am J Clin Nutr 1979; 32: 607-29. 15. Golden MHN, Golden BE, Jackson AA. Method of trace metal balance studies in young children. In: Howell McC J, Gawthorne JM, White CH, eds. Trace element metabolism in animals and man 4. Canberra: Australian Academy of Science, 1981: 69-72. 16. Wilkinson L. Systat: the system for statistics. Evanston, Illinois: Systat, 1990. 17.
Eyre DR, Koob TJ, Van Ness KP. Quantification of hydroxypyridinium crosslinks in collagen by high-performance liquid chromatography.
Anal Biochem 1984; 137: 380-88. Eyre DR, Dickson IR, Van Ness KP. Collagen crosslinking in human bone and cartilage: age related changes in the content of mature hydroxypyridinium residues. Biochem J 1988; 252: 495-500. 19. Beardsworth LJ, Eyre DR, Dickson IR. Changes with age in the urinary excretion of lysyl- and hydroxylysylpyridinoline, two new markers of bone collagen turnover. J Bone Mineral Res 1990; 5: 671-76. 20. Lampl M, Johnston FE, Malcolm LA. The effects of protein supplementation on the growth and skeletal maturation of New 18.
Guinean school children. Ann Hum Biol 1978; 5: 219-27.
21.
Cooper ES, Bundy DAP, MacDonald TT, Golden MHN. Growth suppression in the trichuris dysentry syndrome. Eur J Clin Nutr 1990;
44: 285-91. 22. Heikens GT, Schofield WN, Dawson S, Grantham-McGregor SM. The Kingston project I, growth of malnourished children during rehabilitation in the community, given a high energy supplement. Eur J Clin Nutr 1989; 43: 145-60. 23. Baumgartner RN, Roche AF, Himes JH. Incremental growth tables: supplementary to previously published charts. Am J Clin Nutr 1986; 43: 711-22. 24. Prudhon C, Sall G, Ndiaye B, Lemonnier D. Nutritional regulation of serum osteocalcin: study in kwashiorkor. C R Acad Sci III 1991; 313: 233-38. 25. Trotter M, Peterson RR. Weight of the skeleton during postnatal development. Am J Phys Anthropol 1970; 33: 313-24. 26. Walker SP, Golden MHN. Growth in length of children recovering from severe malnutrition. Eur J Clin Nutr 1988; 42: 395-404.
KH, Black RE, Becker S. Seasonal changes in nutritional status and the prevalence of malnutrition in a longitudinal study of young children in rural Bangladesh. Am J Clin Nutr 1982; 36: 303-13. 28. Golden MHN, Waterlow JC, Picou DIM. The relationship between dietary intake, weight change, nitrogen balance and protein turnover in man. Am J Clin Nutr 1977; 30: 1345-48. 27. Brown
Comparison of saliva and serum for HIV surveillance in developing countries
Saliva has been proposed as a non-invasive alternative to serum for HIV antibody testing. In a field study in Myanmar (formerly Burma), we evaluated such an alternative to identify the frequency of HIV infection in a surveillance programme of high-risk and low-risk sentinel groups. Duplicate vials of saliva and serum were collected from 479 high-risk and 1039 low-risk subjects. One vial of each pair was analysed blind in two laboratories, one in the USA and the other in Myanmar. The US laboratory followed WHO confirmatory strategy II I with three different enzymelinked immunosorbent assays (ELISAs), while the laboratory in Myanmar followed strategy I with one ELISA. Serum testing in the US was the gold standard. The Cambridge ELISA with saliva was a more effective surveillance tool (sensitivity 90·5%, specificity 99·5-100%) for describing the frequency of subjects with HIV antibodies than the serum ELISA supplied to Myanmar by WHO (95·9% and 98·3%, respectively). Saliva is recommended as a safe and effective alternative to serum for HIV antibody testing with ELISA in surveillance programmes in developing countries. ..-
.
v
Lancet 1992; 340: 1496-99.
Introduction Infection with HIV-1 is
becoming evident in Myanmar The (formerly Burma). frequency of HIV-1 infection has risen dramatically in Myanmar to 2-3% during 1990-92. Starting in April-May, 1992, a sentinel surveillance
programme was established, modelled after the successful programme in Thailand.l About 100 subjects are selected every 6 months from high-risk and low-risk groups at various sites. In addition, information is gathered on all blood donors at selected sites. The World Health Organisation (WHO) provides most of the HIV kits but because the number of kits is limited, the sentinel sites and risk groups have to be carefully selected to provide maximum information. The first round of the programme showed that the HIV epidemic is spreading in intravenous drug users, has infected sexually active individuals in middle-risk groups (patients at sexually transmitted disease [STD] clinics, prostitutes) and a low-risk group (women attending antenatal clinics), and is spreading to the general
community (blood donors) (fig 1). If the epidemic in Myanmar is to be controlled, testing needs to be more widespread to make people aware that HIV has reached their community. Yet for Myanmar, as for many developing countries, the cost of widespread testing is too great. Test kits require foreign currency, and personnel to draw blood need to be trained from people who may fear HIV infection from needles. Saliva has been suggested as a non-invasive alternative for HIV testing2-6 in surveillance programmes and in community-based surveys, although in a developing country, climate and conditions are far from ideal for saliva tests.5,6 To
assess
the value of saliva for
surveillance, we did a field study in Myanmar. ADDRESSES: Department of Epidemiology, University of California at Los Angeles (Prof R. R. Frerichs, DrPH); AIDS Prevention and Control Programme, Department of Health, Yangon, Myanmar (M. T. Htoon, MBBS); Saliva Diagnostic Systems, Vancouver, Washington, USA (N. Eskes, BS); and National Health Laboratory, Yangon, Myanmar (S. Lwin, MBBS). Correspondence to Prof R. R. Frerichs, Department of Epidemiology, UCLA School of Public Health, Los Angeles, California 90024-1772, USA.
