ORIGINAL ARTICLE

Effect of Nutritional Status on Tuberculin Skin Testing

Roi Piñeiro & María José Cilleruelo &

Milagros García-Hortelano & Marta García-Ascaso &

Antonio Medina-Claros & María José Mellado

Received: 3 January 2012 /Accepted: 25 April 2012 /Published online: 13 May 2012# Dr. K C Chaudhuri Foundation 2012

AbstractObjectives To evaluate Tuberculin skin test (TST) results ina population of immigrants and internationally adopted chil-dren from several geographical areas; to analyze whethernutritional status can modify TST results.Methods This cross-sectional observational study includedadopted children and immigrants evaluated in the authors’unit between January 2003 and December 2008. Childrendiagnosed with tuberculosis, or vaccinated with live attenu-ated virus 2 mo earlier, HIV-infected, chronically ill orunder treatment with immunosuppressive agents were ex-cluded. TST was considered as dependent variable. Inde-pendent variables were gender, age, geographical origin,BCG scar, nutritional status, immune status and intestinalparasitism.Results One thousand seventy four children were included;69.6 % were girls. There was a BCG scar in 79 % ofchildren. Mantoux00 mm was found in 84.4 %, <10 mmin 4.1 %, and ≥10 mm in 11.4 % of children. Nutrition(McLaren’s classification) was normal (≥90 %) in 26.7 %of the subjects, with mild malnutrition (80–89 %) in 36 %,moderate (70–79 %) in 23.2 % and severe (≤69 %) in14.1 %. There was no difference in TST results amongdifferent nutritional status children.Conclusions The nutritional status, measured by McLaren’sclassification, does not changes the results of TST.

McLaren’s classification only grades protein-caloric malnu-trition, so in authors’ experience this type of malnutritiondoes not interfere with TST results. Implementing othernutritional parameters could help to determine whether nu-tritional status should be taken into account when interpret-ing TST results.

Keywords Tuberculin test . Malnutrition . Child .

Tuberculosis

Introduction

The tuberculin skin test remains as the most useful methodfor the diagnosis of tuberculosis and latent tuberculosisinfection, both in children and adolescents. It is based ondelayed hypersensitivity to certain antigenic components ofMycobacterium tuberculosis [1, 2].

Most of consensus documents concerning tuberculin skintest interpretation agree that malnutrition can cause a certainstate of immunosuppression, and therefore can cause a tuber-culin skin test false negative result [2, 3]. The last Spanishconsensus document on the diagnosis of tuberculosis in pedi-atrics, issued by the Spanish Society of Pediatric InfectiousDiseases, concludes that only some co-morbidities canmodifytuberculin skin test results, without specifying nutritional sta-tus [1]. Scientific publications in this regard, which are de-tailed later, show conflicting results.

Thus the present study was undertaken to evaluate tuber-culin skin test results in a population of immigrants andinternationally adopted children from several geographicalareas and to analyze whether nutritional status can modifytuberculin skin test results.

R. Piñeiro (*) :M. J. Cilleruelo :M. García-Hortelano :M. García-Ascaso :A. Medina-Claros :M. J. MelladoPediatrics Infectious and Tropical Diseases Unit,Department of Pediatrics, Hospital Carlos III,Sinesio Delgado 10,28029 Madrid, Spaine-mail: roipineiro@telefonica.net

Indian J Pediatr (April 2013) 80(4):271–275DOI 10.1007/s12098-012-0773-1

Material and Methods

The Pediatric Infectious and Tropical Diseases Unit inwhich the study was performed is a National ReferenceClinic for the evaluation of internationally adopted childrenand immigrants. These children are requested to get a tuber-culin skin test done upon arrival in Spain, usually when theymake their first clinical visit. The microbiology laboratoryhas a specific parasitology unit employing highly special-ized techniques and a widespread practice of the diagnosisof tropical diseases.

A cross-sectional observational study was performed,which was approved by the ethics committee of the authors’hospital. Patient enrollment commenced in January 2003and finished in December 2008.

Inclusion criteria were: Immigrants or internationally adop-ted children, aged between 0 - 14 y, whose guardians or parentshad requested a health examination at the Pediatric Infectiousand Tropical Diseases Unit; acceptance by guardians orparents, and by adolescents over 12 y, to be included in thestudy after reading and signing an informed consent document.

Exclusion criteria were: Children who did not completethe requested study; children with signs or symptoms oftuberculosis or non-tuberculosis mycobacteria, or havingrecent contact with tuberculosis patients; children who hadreceived any dose of live attenuated virus vaccine during the2 mo prior to the study; HIV-infected children or children withchronic diseases, diagnosed either before or during the studyand children under treatment with immunosuppressive drugs(not including inhaled corticosteroids, non-immunosuppressivedoses of corticosteroids or steroids administered over a periodof less than 1 wk).

The tuberculin skin test result was considered as thedependent variable and was performed by an internationallyapproved technique (intradermal injection of 2 T. U. (0.1 mL)of PPD RT 23). The results were assessed by a pediatri-cian, who was an expert on pediatric infectious diseases,72 h after the injection by measuring the induration diam-eter in millimeters. Only children without known contact withtuberculosis were studied, so any result greater than or equalto 10 mm was considered as a positive tuberculin skin testresult.

The independent variables were gender, date of birth, ageat first consultation in the hospital, country of origin andlength of stay in the country of origin before arrival inSpain, availability of valid immunization cards, nutritionalstatus, immune status according to the percentage of CD4lymphocytes and intestinal parasitic infestation.

McLaren’s nutritional classification was calculated asactual weight X actual height/percentile 50 weight X per-centile 50 height (percentiles according to the World HealthOrganization growth charts). The results are very similar toWaterloo’s nutritional classification.

The authors performed univariate and multivariate (logR) analyses with SPSS 19.0, with statistical significancebeing assumed for values of p<0.05. The bivariate analysiswas tested using Pearson’s χ2 or Fisher’s exact tests or theStudent’s T-test. Unconditional logistic regression analysiswas used in order to include diverse co-variables associatedwith the dependent variable, thus making the autocorrelationmore accurate and controlling any possibly confoundingfactors.

In the Community of Madrid, the initial health examina-tion of internationally adopted and immigrant children isprimarily conducted at two pediatric hospitals. During thestudy period, the distribution of the children was randombetween both hospitals, so it was believed that this did notcause any bias in the origin or characteristics of the patientsincluded in the study.

A limitation of the study was that it not always beingpossible to establish when the Bacille Calmette-Guérin(BCG) vaccine was administered in children with a BCGscar, especially in those without a valid immunization card.However, in most of the origin countries the BCG vaccine isusually administered at birth.

Results

From January 2003 through December 2008, 1,148 immi-grants or internationally adopted children were evaluated atPediatrics Infectious and Tropical Diseases Unit and ofthese, 1,074 were included in the final data analysis afterthe exclusion of 74 children who fulfilled the exclusioncriteria.

Out of the 1,074 children included in the study, 747(69.6 %) were girls. Of the total study population, 373(34.7 %) came from China, 224 (20.8 %) from Latin Amer-ica, 208 (19.4 %) from India and Nepal, 169 (15.7 %) fromEastern Europe and 100 (9.3 %) from Africa. The mostcommon countries of origin were China (34.7 %), India(15.2 %), Russia (9.4 %), Ethiopia (5.5 %) and Bolivia(5 %).

The average age of the children at the time of the studywas 42.61 mo, ranging between 1 and 192 mo with astandard deviation of 40.66 and the median age was24 mo. The average time spent in their country of originwas 39.5 mo, ranging between 1 and 172 mo with a standarddeviation of 36.6. The variable length of stay in their coun-try of origin was very similar to the age at the time of thestudy, since the majority of immigrants or adopted childrenmade their first visit shortly after arriving in Spain. Thecorrelation between both variables was high, with a corre-lation coefficient r00.97 (p<0.001), so for the rest of thestudy the authors used only the variable age at the time ofthe study.

272 Indian J Pediatr (April 2013) 80(4):271–275

Although valid immunization cards showed BCG vac-cine administration in 709 children (66 %), a BCG scar wasfound on physical examination in a total of 848 (79 %). Dueto the unreliability of the immunization cards regarding theBCG vaccine in all the countries, and because 99 % ofvaccinated children develop a scar [4, 5], only the presenceof a BCG scar on physical examination was taken intoaccount when assessing the possible effect of the vaccineon tuberculin skin test results.

The results of Tuberculin Purified Protein Derivative(PPD) reactivity 72 h after the injection were:

& Mantoux00 mm in 907 (84.4 %, 95 % CI: 82.2 %–86.7 %)

& Mantoux <5 mm in 19 (1.8 %, 95 % CI: 0.9 %–2.6 %)& Mantoux 5–9 mm in 25 (2.3 %, 95 % CI 1.4 %–3.3 %)& Mantoux ≥10 mm in 123 (11.4 %, 95 % CI: 9.5 %–

13.4 %)

Chest radiography was performed in all children with atuberculosis skin test result ≥10 mm. The radiographsresults were normal or otherwise these children would havefulfilled an exclusion criterion. All of them were consideredto have latent tuberculosis infection and were treatedaccording to the current recommendations of the SpanishSociety of Pediatric Infectious Diseases [1, 6].

McLaren’s nutritional classification [7] used, and only in287 children (26.7 %) the value was equal to or greater than90 %, indicating a normal nutritional status of children onarrival in Spain. Some degree of malnutrition was found intwo-thirds of the children.

& Mild malnutrition (80–89 %) in 387 children (36 %)& Moderate malnutrition (70–79 %) in 249 children

(23.2 %)& Severe malnutrition (≤69 %) in 151 children (14.1 %)

The relationship between McLaren’s classification andthe geographical area of origin of the children is shown inTable 1. The relationship between nutritional status and thefact that the child was either adopted or immigrant is reflected

in Table 2. The nutritional status of adopted children whenarriving in Spain was worse than immigrant children (p00.003), with a higher percentage of malnourished childrenfrom India, Nepal and Eastern Europe (p00.02).

Two key study groups were defined as PPD0 and PPD10.In the first group, the authors selected those children withTuberculin Purified Protein Derivative reactivity equal to0 mm when compared to any other result, while in thesecond group, children were divided according to the tuber-culin skin test result: less than 10 mm compared to equal toor greater than 10 mm. Different study groups and theircharacteristics are shown in Tables 3 and 4.

In the univariate analysis only two variables, namely ageand the presence of a BCG scar, influenced the tuberculinskin test result. The remaining variables, including nutri-tional status, did not change the result of Tuberculin PurifiedProtein Derivative reactivity in any of the study groups. Theresults of the multivariate analysis in the PPD10 group weresimilar to the findings described above. The only variablesthat influenced the tuberculin skin test result were age andthe presence of a BCG scar.

Discussion

Most of the scientific articles published before 1980report that malnourished children have a lower percent-age of tuberculin skin test positive results when com-pared to children with normal nutrition. In 1968, 51children with severe malnutrition and tuberculosis wereanalyzed [8], and only 18 (35.3 %) had an indurationgreater than 10 mm in the tuberculin skin test. Fourteenyears later, response to the BCG vaccination was reviewedin 261 children from India [9]. A tuberculin skin test wasperformed 6 mo after immunization and nutritional status wasgenerally poor. There were no differences with regard to thetuberculin skin test result in mild degrees of malnutrition,but differences did exist in children with kwashiorkormalnutrition.

Moreover, there are recent publications which considerthat malnutrition could cause a tuberculin skin test falsenegative result. In 2005, Pelly et al. [10] examined 212 adultTable 1 Nutritional status according to McLaren’s nutritional classi-

fication and cumulative percentages by geographical areas

Área/McLaren

<70 % 70–79 % 80–89 % ≥90 %

China 24 (6.4 %) 84 (22.5 %) 155 (41.6 %) 110 (29.5 %)

LatinAmérica

11 (4.9 %) 48 (21.4 %) 91 (40.6 %) 74 (33.1 %)

India andNepal

73 (35.1 %) 53 (25.5 %) 48 (23.1 %) 34 (16.3 %)

EasternEurope

30 (17.8 %) 46 (27.2 %) 58 (34.3 %) 35 (20.7 %)

África 13 (13 %) 18 (18 %) 35 (35 %) 34 (34 %)

Table 2 Nutritional status according to McLaren’s nutritional classi-fication and cumulative percentages by adopted or immigrant childrenand geographical areas

Type/McLaren

<70 % 70–79 % 80–89 % ≥90 %

Adopted 144 (16.8 %) 205 (23.9 %) 299 (34.9 %) 210 (24.4 %)

Immigrant 7 (3.2 %) 44 (20.4 %) 88 (40.7 %) 77 (35.7 %)

Indian J Pediatr (April 2013) 80(4):271–275 273

subjects in Peru and those with protein malnutrition had alower percentage of tuberculin skin test positive results. In2010, another study conducted with 97 Nigerian children[11] indicated that both HIV-infected and underfed subjectswere significantly associated with a tuberculin skin testnegative result. The authors concluded that in these casesthe tuberculin skin test was a slightly reliable technique todiagnose tuberculosis.

However, most of the papers published after 1990, basedon a greater number of children, cast doubt on the classicalrelationship between malnutrition and tuberculin skin testnegative results. In 1995, Verma et al. [12] analyzed 464neonates born in India and BCG-vaccinated at birth. Ofthese, 240 were underweight at birth. The tuberculin skintest was performed at 3, 6 and 36 mo of age and theTuberculin Purified Protein Derivative reactivity was notlower in children with poor nutritional status, although theauthors report that no child suffered from severe malnutri-tion. Four years later, another study evaluated a group of274 Filipino children [13], with 205 having been exposed totuberculosis and 79 being non-exposed. Malnutrition, asmeasured by national growth charts, worsened the clinicalcourse of those who developed tuberculosis, but did notmodify the tuberculin skin test results. Finally, a recentstudy carried out in India examined 3,355 school childrenaged between 5 and 8 y. Weight, height and body massindex, based on the World Hearth Organization growthcharts, were used to define nutritional status. Up to 30 %of children had moderate or severe malnutrition. However,there were no significant differences in tuberculin skin testresults dependent on nutritional status [14].

In the present series, nutritional status was evaluatedaccording to McLaren’s classification [4] and 25.8 % of

the children with an induration equal to 0 mm had a normalnutritional status, mild malnutrition was found in 36.5 %,moderate malnutrition in 23.2 % and severe malnutrition in14.6 %. These percentages were similar among childrenwith any degree of induration (p00.36) and also amongthose with an induration greater than or equal to 10 mm(p00.23).

Conclusions

McLaren’s nutritional classification did not modify the tu-berculin skin test results. Only age and the presence of aBCG scar influenced the tuberculin skin test results in thepresent study. The nutritional classification applied princi-pally grades protein-energy malnutrition, so in authors’experience this type of malnutrition does not modify tuber-culin skin test results. Other methods for the classification ofmalnutrition were not implemented in the authors’ hospitalat the commencement of the study, so it was later decidedto continue employing McLaren’s classification so as toensure a uniform analysis of the data. McLaren’s classi-fication method is still being used at present [15]. It ispossible that the association of McLaren with othermethods such as body mass index, Moore or Waterloo,could be more specific in determining whether nutritionalstatus should be taken into account or not when inter-preting tuberculin skin test results.

Note: The results, discussion and conclusions presentedin this manuscript with respect to malnutrition status areoriginal. The patients, methods and demographic data arepart of a larger study and some of them have been publishedor are ahead-of-print in other scientific journals.

Table 3 Characteristics ofgroups: TST00 mm vs. TST≥1 mm. TST Tuberculin skin test

PPD0 TST00 mm TST≥1 mm pN0907 N0167

Nutrition (McLaren) ≥90 % 234/907 (25.8 %) 53/167 (31.7 %) 0.3680 %–89 % 331/907 (36.5 %) 56/167 (33.5 %)

70 %–79 % 210/907 (23.2 %) 39/167 (23.4 %)

<70 % 132/907 (14.6 %) 19/167 (11.4 %)

Table 4 Characteristics ofgroups: TST<10 mm vs. TST≥10 mm. TST Tuberculin skin test

PPD10 TST<10 mm TST≥10 mm pN0951 N0123

Nutrition (McLaren) ≥90 % 246/951 (25.9 %) 41/123 (33.3 %) 0.2380 %–89 % 348/951 (36.6 %) 39/123 (31.7 %)

70 %–79 % 219/951 (23 %) 30/123 (24.4 %)

<70 % 138/951 (14.5 %) 13/123 (10.6 %)

274 Indian J Pediatr (April 2013) 80(4):271–275

Conflict of Interest None.

Role of Funding Source None.

References

1. Moreno Pérez D, Andrés Martín A, Altet Gómez N, et al. Diagnosisof tuberculosis in paediatrics. An Pediatr (Barc). 2010;73:143.e1–143.14.

2. Piñeiro Pérez R, Cilleruelo Ortega MJ, Mellado Peña MJ. Indica-tions and interpretation of tuberculin skin test. Form Act PediatrAten Prim. 2010;3:114–16.

3. Grupo de Trabajo de Tuberculosis de la Sociedad Española deInfectología Pediátrica. Interpretation of the tuberculin skin testin children. An Esp Pediatr. 2003;59:582–5.

4. Fjällbrant H, Ridell M, Larsson LO. BCG scar and tuberculin reac-tivity in children and adults. Scand J Infect Dis. 2008;40:387–92.

5. Santiago EM, Lawson E, Gillenwater K, et al. A prospective studyof bacillus Calmette-Guérin scar formation and tuberculin skin testreactivity in infants in Lima. Peru Pediatr. 2003;112:e298.

6. Grupo de Trabajo de Tuberculosis de la Sociedad Española deInfectología Pediátrica. Consensus document on treatment of

tuberculosis exposure and latent tuberculosis infection in children.An Pediatr (Barc). 2006;64:59–65.

7. McLaren DS, Read WW. Classification of nutritional status inearly childhood. Lancet. 1972;2:146–8.

8. Lloyd AV. Tuberculin test in children with malnutrition. Br Med J.1968;3:529–31.

9. Satyanarayana K, Bhaskaram P, Seshu VC, Reddy V. Influence ofnutrition on postvaccinial tuberculin sensitivity. Am J Clin Nutr.1980;33:2334–7.

10. Pelly TF, Santillan CF, Gilman RH, et al. Tuberculosis skin testing,anergy and protein malnutrition in Peru. Int J Tuberc Lung Dis.2005;9:977–84.

11. Gofama MM, Garba AM, Mohammed AA. Mantoux test reactionsamong children managed for tuberculosis in Maiduguri, Nigeria.Scand J Infect Dis. 2011;43:15–8.

12. Verma M, Chhatwal J, Aneja R. Tuberculin sensitivity in low birthweight and malnourished children. Indian Pediatr. 1995;32:354–7.

13. Lao LY, De Guia T. Tuberculin skin testing: determinants andreaction. Respirology. 1999;4:311–7.

14. Chadka VK, Jitendra R, Kumar P, Gupta J. Umadevi. Relationshipof nutritional status with tuberculin sensitivity. Indian J Pediatr.2009;76:605–7.

15. Le Grange D, Doyle PM, Swanson SA, Ludwig K, Glunz C,Kreipe RE. Calculation of expected body weight in adolescentswith eating disorders. Pediatrics. 2012;129:e438–46.

Indian J Pediatr (April 2013) 80(4):271–275 275