The prognostic value of dipstick urinalysis in children admitted to hospital with severe malnutrition.

Nahashon Thuo1, Eric Ohuma1, Japhet Karisa1, Alison Talbert1, James A Berkley1, 2

Kathryn Maitland1,3

Affiliations:

1 Centre for Geographic Medicine Research (Coast), Kenya Medical Research Institute,

P. O. Box 230, Kilifi, Kenya

2Centre for Clinical Vaccinology and Tropical Medicine, University of Oxford,

Headington, UK

3Imperial College, London, UK

Keywords:

Severe malnutrition, UTI, Dipstick, Urinalysis, Prognosis

Word count:

Abstract 248

Main text 2508

Author for Correspondence:

Nahashon Thuo

Centre for Geographic Medicine Research (Coast), Kenya Medical Research Institute.

P.O. Box 230, Kilifi, Kenya

Telephone: +254 41 7522063

Fax: +254 41 7522390

Email: nthuo@kilifi.kemri-wellcome.org

Abstract

Background

Children with severe malnutrition (SM) present to hospital with an array of

complications, resulting in high mortality despite adherence to World Health

Organization guidelines. Diagnostic resources in developing countries are limited and

bedside tests would help identify high-risk children. Dipstick urinalysis is a bedside

screening test for urinary tract infections (UTIs). UTIs are common in SM and can lead

to secondary invasive bacterial sepsis. Very few studies have examined the usefulness

of dipstick screening of urine specimen in SM and none has explored its prognostic

value.

Patients and Methods

A 2 year prospective study on children admitted in Kilifi district hospital, Kenya, with SM.

Freshly voided, clean catch urine samples were tested using Multistix reagent test

strips, and positive samples sent for culture.

Results

Out of the 667 children admitted, 498 children (75%) provided urine samples; of these,

119 (24%) were positive for either leucocyte esterase (LE) or nitrites. Twenty eight

children (6% overall) had UTI confirmed by urine culture. All isolates were coliforms ;>

50% were resistant to cotrimoxazole and gentamicin. There was no difference in signs

of severity between those with positive urine dipstick and those without. Case fatality

was higher among children with a positive dipstick (29% vs. 12%). Presence of a

positive dipstick was a strong predictor of mortality (adjusted HR: 2.76)

Conclusions

A urine dipstick positive for either LE or nitrites is a useful predictor of death in children

admitted with SM and can help guide antimicrobial treatment.

Introduction

Children with severe malnutrition (SM) presenting to hospital often have an array of

complications. As a consequence, management with a single protocol is challenging

and may not be optimal for all groups. In particular, children with SM are at risk of

infection(1, 2); current guidelines recommend parenteral antibiotics for cases with

complicated disease(3). Sepsis has been identified as a major contributor of death in

children with SM and is associated with both early and late case fatality(4, 5). Because

of limited diagnostic resources and facilities in developing countries, rapid tests to

identify high-risk children are needed to enable supportive interventions to be targeted

effectively and improve current management guidelines.

Whilst urinary tract infections (UTIs) are said to be a common complication in SM

reported prevalence varies between 3% and 35% among paediatric admissions with

malnutrition(6-8). UTIs are caused predominantly by Gram-negative enteric bacilli. The

increased susceptibility of malnourished children to UTIs is thought to be due to their

transient state of immunodeficiency characterized by breakdown of anatomical barriers,

decreased cell-mediated immunity, decreased phagocytosis and opsonization. (9-11). It

is postulated that these factors allow organisms to ascend up the urinary tract, and may

lead to secondary invasive bacterial sepsis. The World Health Organization

recommends urine microscopy and culture to diagnose UTI(3). Urine culture is not

practical as it takes at least 48 hours to give a result whilst microscopic examination of

urine is time consuming and labor intensive. Moreover, with a deliberate shift to

community-based management of severe malnutrition(12), performing microscopy and

culture is not practical.

Dipstick urinalysis is an excellent, low-cost, bedside screening test in children. It is used

to screen for urinary tract infections (nitrites and leucocyte esterase), diabetes mellitus

(glucose), nephrotic syndrome (protein), glomeluronephritis (blood) to guide diagnostic

work-ups and for assessment of hydration status (specific gravity). Reagent strips

perform as well as microscopy in the diagnosis of urinary tract infection in many patient

groups, including children (13, 14) . Despite concerns of the specificity and thus

diagnostic accuracy, studies have shown that a strategy that combines use of nitrites

and leucocyte esterase (LE) testing appears to offer the best performance(15). Nitrites

(a bacterial metabolite of dietary nitrate) and leucocyte esterase (from white cells) are

not normally present in sterile urine. A negative test gives a negative predictive value

(NPV) of 97% (specificity 98.7%)(14, 16, 17). Positive results identify the group

requiring urine culture and may help target first-line treatment.

From the studies in developing countries that have looked at UTI in malnourished

children, none have examined the usefulness of dipstick screening of urine specimens

or explored its prognostic value. In this study we sought to establish if the presence of

LE and nitrites in urine correlates with disease severity and fatal outcome in

malnourished children.

PATIENTS AND METHODS

Study site

A large prospective, observational study conducted at Kilifi District Hospital paediatric

ward at the Centre for Geographical Medicine Research, on the coast of Kenya

between June 2005 and June 2007. Ethical approval was granted by the Kenya

Medical Research Institute (KEMRI) Scientific Steering Committee and the National

Ethics Review Committee (SSC No 927).

Patients and consent

All children referred to the general paediatric ward for admission were assessed for

eligibility. At admission, the admitting clinician completed a standardised medical case

report form documenting medical history and physical examination. Ward assistants

were trained to take anthropometric measurements (height, weight and mid-upper arm

circumference) on all children. Following an explanation of the study in the local

language, written consent was sought from parents or guardians. Children received

routine standard of care specified by the WHO guidelines(3). This included parenteral

antibiotics (ampicillin and gentamicin), a broad-spectrum anti-helminth (mebendazole),

therapeutic milk (F75 initially then F100) and vitamin and mineral supplementation.

Urine testing

Ward assistants were trained to collect urine samples. Urine was collected as soon as

possible after admission through clean catch method. Dipstick testing was done on a

freshly voided urine sample using urinalysis reagent strips (Mission™, Acon

Laboratories Inc, San Diego, USA). If the urine was positive for leukocytes or nitrites, a

further clean catch urine sample was sent for microscopy, culture and sensitivities. This

two stage method has been shown to reduce the possibility of false positive results by

contamination(18). Only urine samples collected on the day of admission were reported.

Laboratory methods

Urine was cultured on CLED agar at 370C. A positive culture was defined as growth of a

single urinary tract pathogen at >= 50 CFU/µL. As part of routine clinical practice, blood

cultures were taken at admission and incubated in a BACTEC 9050 system instrument

(Becton Dickinson, http://www.bd.com). Sensitivities to antimicrobials were investigated

in accordance with the recommendations of the British Society for Antimicrobial

Chemotherapy (BSAC). All the clinical and laboratory data was maintained in a

database (FileMaker 5.5v1, FileMaker Inc, www.filemaker.com).

Definitions

Severe malnutrition was defined as one of: oedema of both feet (of kwashiorkor or

marasmic kwashiorkor) or weight for height Z score ≤ -3 or mid upper arm

circumference (MUAC) < 11cm (if length > 65cm)(19). A positive dipstick was defined

as either leucocyte esterase ≥ (+/-) or nitrites ≥ (+).

Statistical Methods

We categorised children into two groups for analysis; LE or Nitrites positive (positive

dipstick) and both LE and Nitrite negative (negative dipstick). Measures of association

between the groups were evaluated using a Pearson’s chi-square test of association

and Fisher’s exact for small samples. Comparison of means was done using the

unpaired Students t-test for continuous data. We performed survival analysis to

determine the probability of survival in days following admission and assessed the

difference in the survival functions using the logrank test. Both known and potential risk

factors for poor outcome in SM were evaluated by fitting the Cox proportional hazards

model assuming that the baseline force of mortality remains constant over the entire

study period in the two groups. Statistical significance was assessed at the 5% level

with p-values < 0.05 considered to be statistically significant.

RESULTS

Patient Characteristics

Six hundred sixty seven children (313 females) were included in the study. The median

age was 22 months (IQ range 16-35 months). Two hundred and thirty nine (36%) had

oedema. Of all children recruited, 498(75%) gave a urine sample at admission, and

before parenteral antibiotics. Out of 667 children enrolled into the study, 125 children

(19%) died in hospital during the two year study period.

Dipstick results

Of the 498 admission urine sample obtained, 88(18%) were LE positive, 53(11%)

nitrites positive, 111(23%) protein positive and 24(5%) blood positive. Only 22(4%) were

positive for both LE and nitrites while 119(24%) positive for either. A higher proportion of

girls had nitrites positive (14% vs. 8%, P=0.02), leucocytes positive (26% vs. 11%,

P<0.01) and blood positive (7% vs. 3%, P =0.02) samples compared to boys but no

significant differences in protein. Only protenuria showed a difference in frequency by

age, being more common in the infants compared to those older than 1 year (44% vs.

21%, P< 0.001).

Urine culture of 119 samples yielded 28 organisms, all of them coliforms. Of the

isolates, 26(93%) were resistant to cotrimoxazole, 12(43%) to gentamicin, 4(14%) to

nalidixic acid and 6(21%) to nitrofurantoin. Blood culture yielded E. coli from one patient,

contaminants in four while the remaining 23 had no isolate. Resistance to the current

standard treatment was not related to age, gender or outcome (death). There was no

association between a positive urine dipstick and bacteraemia (5% vs. 7%, P= 0.54).

Case fatality Children with no urine sample collected on admission were more severely ill and had

greater case fatality (Table 1).They were more likely to have features of severe disease

namely shock, impaired consciousness and bacteraemia with nine (20%) of these

children died within 48 hours of admission. Forty two percent of those who died had a

positive urine dipstick. In-patient fatality was 30(26%), 8(34%), 14(26%) and 25(28%)

for protein, blood, nitrite and LE respectively. Children with LE or Nitrite positive had a

higher fatality compared to those with negative dipstick (29% Vs 12%, P<0.01). The

proportions of children with features of severe disease at admission were compared

between the two groups (dipstick positive and dipstick negative) but did not show any

significant differences (Table 2) except for hypokalemia. There was no association

between a positive urine culture and death (19% vs. 23%, P= 0.57).

The in-hospital survival functions (probability of survival) were compared in dipstick

positive and dipstick negative children which demonstrated a significant difference

between the two groups (P<0.01). Cox analysis also determined that a positive urine

dipstick was independently associated with death, even after consideration of baseline

variables (HIV status, level of consciousness, dehydration, impaired perfusion,

electrolyte imbalance and bacteraemia) known to affect outcome in SM (table 3).

Children who had a positive dipstick results had nearly three times higher risk of dying

than those who had a negative dipstick result (HR = 2.76 95% CI: 1.62 -4.69, P < 0.01).

The risk of dying was 2.3(95% CI: 1.31-4.09, P= 0.003) and 1.8(95% CI: 1.10-4.50, P=

0.067) for children with LE and nitrites respectively. There was no difference in the

median time to death between the two groups (7.5 days vs. 6days, p =0.627).

DISCUSSION

In this study, approximately a quarter of the 498 children tested had a positive urine

dipstick. Children who had a positive dipstick results were approximately three times

more at risk of dying than those who had a negative dipstick result even after adjusting

for known features of severe disease: bacteraemia, shock, electrolyte imbalance and

impaired consciousness(5, 20, 21). Forty two percent of all children who died were

dipstick positive. Overall, 6% of the children had culture proven UTI, with enteric

coliforms being the predominant organism isolated. Most of the isolated organisms had

poor sensitivity to antibiotics commonly used in malnutrition; 40% were resistant to

gentamicin and 90% resistant to cotrimoxazole.

Our findings of high resistance patterns to the current standard of care for SM confirms

previously reported low sensitivity to first line antibiotics (7, 22). While this could have

been due to prior use of antibiotics as have been reported previously (23), these findings

are worrying . Poor sensitivity to antibiotics commonly used in malnutrition may

contribute to both primary failure and later failures due to recrudescence of inadequately

treated pathogens. This supports the need for surveillance of antimicrobial resistance

patterns to advice on applicable recommendations on antimicrobial use in UTIs.

To the best of our knowledge, no previously published study has described mortality

according to dipstick finding. We could not find any association between urine dipstick

and bacteraemia but blood cultures are known to be insensitive for detecting bacteremia.

We postulate that since malnourished children have relative immunosuppresion due to

nutritional factors and/or active immunosuppresion due to HIV or TB, they are at risk

invasive bacterial sepsis secondary to inadequately treated pathogens. The high

mortality and poor sensitivity to routine antibiotics in the background relative

immunosuppresion prompts the need for clinical trials to determine if changes in the

current first line therapy would improve outcome in children with UTI or in whom UTI is

suspected (e.g. positive urine dipstick).

The choice of routine antibiotics should be guided not only by the susceptibility of likely

pathogens, but also by the sites of infection, the ability of the antibiotics to penetrate the

sites and immune response of the host. Children with severe malnutrition are susceptible

to both skin and urinary tract infection suggesting the need for antibiotics with good tissue

penetration and good renal excretion. Though combinations of beta-lactams and

aminoglycosides are excellent first line therapy, consideration should be given to

fluoroquinolones since they possess an enlarged antimicrobial spectrum, greatly

enhanced bactericidal activity, and substantial pharmacokinetic advantages compared

to nalidixic acid. Due to safety concerns and limited published experience of their use,

quinolones use in children is limited to life-threatening or difficult to treat infection and

circumstances where other antibacterial agents cannot be used. (24, 25). Results of

published clinical trials with fluoroquinolones in pediatric patients show promising

efficacy and safety and may be of benefit to children with severe malnutrition (4) but

pharmacokinetic studies are required first.

One of the limitations of our study was the challenges we encountered when collecting

urine in severely ill children. We may have underestimated the true prevalence of UTI

and prognostic value of urine dipstick since we were unable to collect urine in a quarter

of the children. Most of these cases were severely ill, either in shock or had depressed

level of consciousness. Catheter urine specimen has been used in pediatric ICU setting.

However it is expensive and always poses the risk of introducing infection. Urine

collection by means of adhesive perineal bag is a widely used method in children who

cannot control urine emission. It is cheap and easy to use(26). However, this technique

has a high risk of contamination and a very low positive predictive value(27). In poor

resource setting it would be worthwhile to evaluate its utility when proper cleaning of the

perineal area before urine collection is done as this has been shown to reduce the

contamination rate(28). Our interpretation of the data was also limited by selective urine

culture and prior antibiotic use which may underestimate prevalence of UTIs and

increase antimicrobial resistance.

Given the high mortality observed in malnourished children especially those with

bacteraemia, there is need to identify simple and robust technologies for rapid diagnosis

of the potential foci of infection. An ideal screening test should be inexpensive, easily

accessible and simple to do. In this regard, dipstick urinalysis fares well. A single

reagent strip costs $0.15 compared to $4 for urine microscopy. Although not very

specific in diagnosing UTIs, it identifies children at a very high risk of dying and is thus a

useful in SM.

Conclusion

The results of the dipstick urinalysis suggest that they can be utilized as a valuable

prognostic marker among children admitted to hospital with SM. Use of dipstick

urinalysis in SM will help identify children at high risk and may help in deciding

appropriate treatment. Clinical trial are needed to determine appropriate first line therapy

in children with SM.

Author Contributions

NT, AT and JK provided inpatient care and data collection. EO, JB and NT conducted

analysis and preparation of the manuscript for submission. KM conceived and designed

the study, conducted the statistical analysis and the overall manuscript preparation. All

authors contributed to the final manuscript.

Acknowledgements

We are grateful to the subjects for their participation in the study. We are also grateful to

all members of the KEMRI laboratory and computing team who participated in data

collection and data storage. The study was supported by the Kenya Medical Research

Institute and the Wellcome Trust. JB is supported by a fellowship from the Wellcome

Trust. This paper is published with the permission of the Director of the Kenya Medical

Research Institute (KEMRI).

Conflicts of interests

The authors have no potential conflict of interest to declare.

Role of the funding source

The sponsor of the study had no role in study design, data collection, data interpretation

or writing of the report.

WHAT IS ALREADY KNOWN ON THIS TOPIC

Severe malnutrition is associated with high mortality; however there are very few rapid

tests that can identify high risk children. Dipstick urinalysis is used as a screening test

for UTIs in children, a condition that is common in severely malnourished children.

WHAT THIS STUDY ADDS

Dipstick urinalysis in children with severe malnutrition identifies children at high risk of

dying and can be used to stratify patients for interventions with antibiotics.

References

1. Berkowitz FE. Infections in children with severe protein-energy malnutrition. Annals of tropical paediatrics. 1983 Jun;3(2):79-83.

2. Sunguya BF, Koola JI, Atkinson S. Infections associated with severe malnutrition among hospitalised children in East Africa. Tanzania health research bulletin. 2006 Sep;8(3):189-92.

3. WHO. Management of Severe Malnutrition: A Manual For Physicians And Other Senior Health Workers; 1999.

4. Babirekere-Iriso E, Musoke P, Kekitiinwa A. Bacteraemia in severely malnourished children in an HIV-endemic setting. Annals of tropical paediatrics. 2006 Dec;26(4):319-28.

5. Maitland K, Berkley JA, Shebbe M, Peshu N, English M, Newton CR. Children with severe malnutrition: can those at highest risk of death be identified with the WHO protocol? PLoS medicine. 2006 Dec;3(12):e500.

6. Bagga A, Tripathi P, Jatana V, Hari P, Kapil A, Srivastava RN, et al. Bacteriuria and urinary tract infections in malnourished children. Pediatric nephrology (Berlin, Germany). 2003 Apr;18(4):366-70.

7. Rabasa AI, Shattima D. Urinary tract infection in severely malnourished children at the University of Maiduguri Teaching Hospital. Journal of tropical pediatrics. 2002 Dec;48(6):359-61.

8. Reed RP, Wegerhoff FO. Urinary tract infection in malnourished rural African children. Annals of tropical paediatrics. 1995;15(1):21-6.

9. Chandra RK. 1990 McCollum Award lecture. Nutrition and immunity: lessons from the past and new insights into the future. The American journal of clinical nutrition. 1991 May;53(5):1087-101.

10. Rikimaru T, Taniguchi K, Yartey JE, Kennedy DO, Nkrumah FK. Humoral and cell-mediated immunity in malnourished children in Ghana. European journal of clinical nutrition. 1998 May;52(5):344-50.

11. Katona P, Katona-Apte J. The interaction between nutrition and infection. Clin Infect Dis. 2008 May 15;46(10):1582-8.

12. World Health Organization tWFP, the United Nations System Standing Committee on Nutrition and the United Nations Children’s Fund. COMMUNITY-BASED MANAGEMENT OF SEVERE ACUTE MALNUTRITION. 2007 [cited 2008 10/12/2008]; Available from: http://www.who.int/nutrition/topics/Statement_community_based_man_sev_acute_mal_eng.pdf

13. Hiraoka M, Hida Y, Hori C, Tuchida S, Kuroda M, Sudo M. Rapid dipstick test for diagnosis of urinary tract infection. Acta paediatrica Japonica; Overseas edition. 1994 Aug;36(4):379-82.

14. Gorelick MH, Shaw KN. Screening tests for urinary tract infection in children: A meta-analysis. Pediatrics. 1999 Nov;104(5):e54.

15. Raymond J, Sauvestre C. [Microbiological diagnosis of urinary tract infections in the child. Importance of rapid tests]. Arch Pediatr. 1998;5 Suppl 3:260S-5S.

16. Hurlbut TA, 3rd, Littenberg B. The diagnostic accuracy of rapid dipstick tests to predict urinary tract infection. American journal of clinical pathology. 1991 Nov;96(5):582-8.

17. Whiting P, Westwood M, Watt I, Cooper J, Kleijnen J. Rapid tests and urine sampling techniques for the diagnosis of urinary tract infection (UTI) in children under five years: a systematic review. BMC pediatrics. 2005;5(1):4.

18. Loane V. Obtaining urine for culture from non-potty-trained children. Paediatric nursing. 2005 Nov;17(9):39-42.

19. Berkley J, Mwangi I, Griffiths K, Ahmed I, Mithwani S, English M, et al. Assessment of severe malnutrition among hospitalized children in rural Kenya: comparison of weight for height and mid upper arm circumference. Jama. 2005 Aug 3;294(5):591-7.

20. Erinoso HO, Akinbami FO, Akinyinka OO. Prognostic factors in severely malnourished hospitalized Nigerian children. Anthropometric and biochemical factors. Tropical and geographical medicine. 1993;45(6):290-3.

21. Tolboom JJ R-MA, Kabir H, Molatseli P, Anderson J. Severe protein energy malnutrition in Lesotho, death and survival in hospital, clinical findings. Tropical and geographical medicine. 1986;38(4):351-8.

22. Caksen H, Cesur Y, Uner A, Arslan S, Sar S, Celebi V, et al. Urinary tract infection and antibiotic susceptibility in malnourished children. International urology and nephrology. 2000;32(2):245-7.

23. Berkley JA, Lowe BS, Mwangi I, Williams T, Bauni E, Mwarumba S, et al. Bacteremia among children admitted to a rural hospital in Kenya. The New England journal of medicine. 2005 Jan 6;352(1):39-47.

24. Koyle MA, Barqawi A, Wild J, Passamaneck M, Furness PD, 3rd. Pediatric urinary tract infections: the role of fluoroquinolones. The Pediatric infectious disease journal. 2003 Dec;22(12):1133-7.

25. The use of systemic fluoroquinolones. Pediatrics. 2006 Sep;118(3):1287-92.

26. Whiting P. Clinical effectiveness and cost-effectiveness of tests for the diagnosis and investigation of urinary tract infection in children: a systematic review and economic model. Health Technology Assessment. 2006;10(36).

27. Alam MT, Coulter JB, Pacheco J, Correia JB, Ribeiro MG, Coelho MF, et al. Comparison of urine contamination rates using three different methods of collection: clean-catch, cotton wool pad and urine bag. Annals of tropical paediatrics. 2005 Mar;25(1):29-34.

28. Vaillancourt S MD, Zhang X. Cleaning of the perineal/genital area before urine collection from toilet-trained children prevented sample contamination. Pediatrics. 2007;119:1288–93.

29. American College of Chest Physicians/Society of Critical Care Medicine Consensus Conference: definitions for sepsis and organ failure and guidelines for the use of innovative therapies in sepsis. Critical care medicine. 1992 Jun;20(6):864-74.

Table 1: Clinical and laboratory features of those who provided an admission urine sample and those who did not.

Variable Dipstick (%) 1 No Dipstick (%) P valuePatients recruited, n 498 169Female 227(46) 86(51) 0.26Age – Median (IQR) 22.4(17-36) 20.9(14-31) 0.23Oedema 178(36) 61(36) 0.97Clinical signs Fever(temperature >38.5C) 197(40) 72(42) 0.52Hypothermia (temperature <36.0C.) 0 2(1) 0.06Tachypnoea4 115(23) 51(30) 0.07Tachycardia 3 137/497(28) 53(31) 0.37Hypoxia (oxygen saturations <90%) 73/497(15) 36(21) 0.05Impaired consciousness2 11(2) 17(10) <0.001Severe anaemia (Hb <5 g/dl) 160(32) 52(31) 0.08Severe dehydration 101(20) 41(24) 0.29Impaired perfusion6 76(15) 44(26) 0.01SIRS5 220(44) 95(56) 0.01Laboratory featuresHyponatremia (sodium<130mmol/l) 236/429(55) 82/146(56) 0.81Hypokalemia (potassium<3.0mmol/l) 155/429(36) 44/146(30) 0.19Bacteraemia 31(6) 21(12) 0.01Leucocytosis(WBC count > 12000/mm3) 461/489(94) 159/164(94) 0.86HIV antibody positive 107/478(22) 36/162(21) 0.62Median days in stay(IQR) 9 (7-14) 10 (7-14) 0.26Died 81(16) 44(26) <0.01Died within 48hrs 6(1) 9(5) <0.011Values in parentheses are percentages.2Impaired consciousness =prostration or coma.3Tachycardia = heart rate > 180, 140, 130 bpm for ages < 12m, 1-5 y, > 5 y respectively. 4Tachypnea =respiratory rate above 50, 40 or 30 bpm for ages < 12m, 1-5 y, > 5 y respectively. 5Severe inflammatory response syndrome(SIRS)= at least two of the following: Core temperature of >38.5C or <36.0C; or tachycardia ; or tachypnea; or leucocytosis, or white blood cell count less than 4,000/cu mm(29)6Impaired perfusion: any one of the following; Capillary refill time >2 sec or temperature gradient or weak pulse volume

Table 2: Selected features of severe disease in dipstick positive and dipstick negative children

Clinical variables Dipstick Positive 1

Dipstick Negative

P value

No of patients, n 119 379Female 72(61) 155(41) <0.001Male 46(39) 225(59)

Median age (IQR) 21(15-31) 23(16-36) 0.14

Oedema 50(42) 128(34) 0.10HIV antibody positive 25/113(22) 90/365(25) 0.52

Clinical signs Fever (temperature >38.5C) 45(38) 152(40) 0.66Tachypnoea 4 24(20) 90(24) 0.37

Tachycardia 3 26/118(22) 111/379(29) 0.12Hypoxia (oxygen saturations <90%) 19/118(16) 54/379(14) 0.62

Impaired consciousness2 3(2) 8(2) 0.73Severe dehydration 29(24) 72(19) 0.20

Severe anaemia(Hb <5 g/dl) 33(28) 127(32) 0.24Impaired perfusion5 16(13) 60(16) 0.53

SIRS 48(40) 172(45) 0.33Laboratory featuresHyponatremia (sodium<130mmol/l) 64/104(60) 172/325(53) 0.12Hypokalemia (potassium<3.0mmol/l) 48/102(47) 107/327(33) <0.01Bacteraemia 6(5) 25(6) 0.67Leucocytosis(WBC count > 12000/mm3) 113/118(96) 348/371(94) 0.50Died (%) 34(29) 47(12) <0.001Died within 48hrs 3(2) 3(1) 0.15

1Values in parentheses are percentages.2Impaired consciousness =prostration or coma.3Tachycardia = heart rate > 180, 140, 130 bpm for ages < 12m, 1-5 y, > 5 y respectively. 4Tachypnea =respiratory rate above 50, 40 or 30 bpm for ages < 12m, 1-5 y, > 5 y respectively. 5Severe inflammatory response syndrome(SIRS)= at least two of the following: Core temperature of >38.5C or <36.0C; or tachycardia ; or tachypnea; or leucocytosis, or white blood cell count less than 4,000/cu mm(29)6Impaired perfusion: any one of the following; Capillary refill time >2 sec or temperature gradient or weak pulse volume

Table 4. Hazard Ratio for death in Children with Malnutrition according to dipstick results

Dipstick Results Odds Ratio (95% Confidence None 1.0Any 2.0(1.1-3.6)LE 2.0(1.0-4.5)Nitrates 1.5(0.7-3.2)Blood 1.7(0.7-4.7)Protein 1.5(0.8-2.8)LE or Nitrates 2.5(1.3-4.5)LE or Blood 1.8(0.9-3.30*Odds ratios are for death, adjusted for age, sex, HIV infection, severity and type of malnutrition, electrolyte imbalance and dehydration.

Figure 1: cumulative hazard curve grouped by dipstick results

0.00

0.10

0.20

0.30

cum

mul

ativ

e ha

zard

0 10 20 30 40Time from admission(days)

Dipstick positive Dipstick negative

Cumulative hazard curve grouped by dipstick results

* The Hazard ratio has been adjusted for sex, age, HIV status, level of consciousness, dehydration, impaired perfusion, hyponatremia, hypokalemia and bacteraemia.

Logrank P value<.001Adjusted* HR 2.76(95% CI 1.62-4.69)