Search tips
Search criteria 


Logo of nihpaAbout Author manuscriptsSubmit a manuscriptHHS Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
Int J Paediatr Dent. Author manuscript; available in PMC 2010 November 1.
Published in final edited form as:
PMCID: PMC2866123

Protein-energy malnutrition during early childhood and periodontal disease in the permanent dentition of Haitian adolescents aged 12–19 years: a retrospective cohort study



The aim of this retrospective cohort study was to examine whether exposure to early childhood protein-energy malnutrition (ECPEM) is related to worsened periodontal status in the permanent dentition during adolescence.


A trained clinician/researcher examined the periodontal status of 96 persons aged 12–19 living in rural Haiti using WHO diagnostic criteria (Community Periodontal Index, WHO 1997). Malnutrition data of the study participants had been collected during the years 1988–1993 by a nongovernmental organization. We compared those who had been malnourished in early childhood, based on z-scores for anthropomorphic data collected during the first 5 years of life, with those who had not been malnourished, regarding mean Community Periodontal Index (CPI) score, controlling for age, sex, socioeconomic status, and smoking.


Overall, 57.3% of the participants demonstrated a CPI score of 3 or greater in at least one sextant. ECPEM was independently and positively related to mean CPI score, when controlling for sex and smoking.


More than half of these young Haitians demonstrated CPI scores of 3 or greater, and ECPEM was related to poorer periodontal status, as measured by CPI, in the permanent dentition.


Data on the relationship between malnutrition and oral outcomes are scarce. The few epidemiologic studies conducted over the past four decades have tended to focus malnutrition effects on the dentition, including effects on tooth eruption patterns1, enamel hypoplasia2, and on dental caries prevalence24. Although epidemiologic studies of acute periodontal conditions in malnourished populations have shown that malnutrition is related to noma (cancrum oris)5,6, and acute necrotizing ulcerative gingivitis7, and that various nutritional deficiencies, including those of calcium8, vitamin D9, ascorbic acid10, and protein11 may be related to various forms of periodontal disease or manifest oral symptoms which mirror periodontal disease (e.g., scurvy), these investigations have all been cross-sectional in nature, have tended to evaluate only the degree of gingivitis but not periodontitis, or have examined the effects of malnutrition on the tooth supporting structures in only the primary dentition.

Protein-energy malnutrition (PEM) occurs when there are deficiencies in protein, energy foods or both, relative to a body’s needs11. PEM during early childhood is thought to have profound effects on the developing immune system12, most likely due to the concomitant occurrence of micronutrient deficiency, and therefore may have effects on susceptibility to infectious disease that last long past early childhood13.

This is a retrospective cohort study designed to evaluate the effects of early childhood protein-energy malnutrition (ECPEM) measured during the first 5 years of life on the periodontal status of the permanent dentition in Haitian adolescents 7–12 years later, when aged 12–19, as measured with the Community Periodontal Index (CPI)14. Data relating to the malnutrition status of the participants during the first 5 years of life were obtained in these participants from the Haitian Health Foundation (HHF); a US-based nonprofit oral re-hydration and nutrition health programme operating in Jeremie, the third largest city of Haiti located in the southern region, and the surrounding rural areas. Since 1988, the HHF has provided a primary health care outreach program for over 120,000 residents of the Jeremie region of Haiti. As part of this health programme, the HHF established, and has continued to maintain, a computerized database on persons served; this database includes monthly weightings for the first 6 months of life and quarterly through 5 years of age for all children born since 1988. The presence of this unique database has allowed us the opportunity to answer important and intriguing scientific questions on the physiology and pathology of oral disease, including caries, tooth eruption patterns, and periodontal disease as they relate to PEM as experienced during early childhood.

The purpose of this retrospective cohort study was to examine whether those adolescents who had been exposed to ECPEM during their first 5 years of life were more likely to subsequently develop periodontal disease by age 12–19, compared to those who were not malnourished.

Materials and methods

Study sample

The sample of participants in this study was drawn from the population of a larger, ongoing investigation that investigated the effects of ECPEM during the first 5 years of life on the permanent dentition of rural Haitian children and adolescents. This larger core investigation included over 1000 Haitians who were enrolled from a total of 19 local villages. Details about the sampling and recruitment methodology for the core study have been previously published1. In order to be eligible for this and for the larger, ongoing study, participants had to have two recorded weighing per year for at least 3 of first 5 years of life in the HHF database and to reside in one of targeted villages in the Jeremie region during the study field period (May–August 2004). This time period was chosen as it guaranteed access to the greatest number of children, as the small per cent of children from Jeremie who are sent to Port-au-Prince for their education return for the summer months as schools close from June to October. Additionally, these months predate the rainy season and its associated travel difficulties.

For this sub-study, a study recruiter invited all eligible children of three villages to participate, and the study staff obtained informed consent, in Creole, from the parent and assent from the corresponding child. The protocol for this study was reviewed and approved by the Institutional Review Board of the New York University School of Medicine.


Early childhood protein-energy malnutrition was based upon height and weight measurements made during the first 5 years of each participant’s life, which were extracted from the database maintained by the HHF. The weighing scales were UNICEF-issued and calibrated at the beginning and during the weighing process. Monthly weightings for the first 6 months of life and quarterly through 5 years of age were the targets for the 1988–1996 period.

All definitions used a z-score based upon the weight-for-age for each child on a given visit, normalized with National Center for Health Statistics (NCHS) data. The NCHS database was selected for normalization to allow international comparisons as recommended by the WHO. A z-score of 0.00 indicates a weight-for-age equal to that of an average child of the same gender in the NCHS database, whereas a score less than 0.00 indicates a weight-for-age that is the indicated number of standard deviations below the average child of the same gender and age in the NCHS database. For this sub-study, ECPEM was defined as a child having any recorded weight-for-age z-score of less than or equal to ≤ − 2.0.

Periodontal status was recorded by a single, trained dentist (G.J.C.). All children examined had a cardiac auscultation prior to periodontal probing; no children were found to have cardiac murmurs. The examinations were conducted in local villages and consisted of a visual dental periodontal examination following World Health Organization Criteria (WHO 1997) for periodontal disease measures. The examiner was standardized to WHO criteria for the Community Periodontal Index Treatment Needs by an experienced examiner prior to the initiation of the data collection phase (field operation), and in the field the examiner was assisted by a trained recorder. As recommended by the WHO, examinations of these children were limited to six index teeth: 16, 11, 26, 36, 31, and 46. We used a colour-coded pressure-sensitive probe (Vivacare; Vivadent, Schaan, Liechtenstein), which was specially designed to meet international standards for the screening of periodontal disease. Periodontal probing measurements were taken at two sites (mesiobuccal and mesiolingual) on each first molar and at two sites (mesiofacial and distofacial) on each of the selected incisor teeth for all participants. A total number of 12 sites per participant (two sites per tooth) received CPI scores ranging from 0, which signified healthy tissue, to a score of 4, which indicated a pathologic pocket. The WHO recommends that, in order to avoid scoring the deepened sulci associated with eruption (specifically of the premolar teeth) as periodontal pockets, scores of 3 and or 4, which indicate pocketing of 4–5 mm and 6+ mm, respectively, not be recorded in children under the age of 1514. Because (i) a substantial proportion (47%) of those examined were aged 12–14 and (ii) periodontal disease is known to occur in children as young as 10–12 years15, we chose to include measurements made on the mesial aspects of the first molars of those younger than age 15. But, in order to examine whether deepened probings in those aged 12–14 (that might be associated with tooth eruption and not with periodontitis per se) accounted for differences found in our primary analysis of the ECPEM–CPI relationship, we conducted additional analyses which included alternative measurements taken from the first molars which would not likely be affected by adjacent erupting premolar teeth (i.e., the direct buccal and direct lingual measurements). The level of periodontal disease was based upon the highest score recorded per tooth; scores were then averaged to produce a mean CPI score per individual.

Other measures

Socioeconomic status (SES) at baseline was taken from the original HHF data set and represents SES status of the participant’s family at birth. The SES questionnaire was developed to define rural Haitian SES status and is based upon the standard Haitian government census form. It has been translated and used in research in rural Haiti16. The goal of the SES questionnaire was to ascertain the children’s household SES for rural Haiti and therefore questions includes items including number of rooms of the house, presence of a working radio, type of roof (e.g., tin, thatch), type of floor, water source, and presence of a latrine. Data may be considered as defining a relative socioeconomic measure. As part of this study, participants were also asked whether or not they presently smoked cigarettes, and were classified as either smokers or nonsmokers. Sex and age were also recorded.

Data management/analysis

All data were double-entered into a computerized database, confirmed against the original datasheets and then converted to an SPSS data set. Frequency cleaning was carried out to correct any possible range or distribution errors. Descriptive and exploratory analyses were conducted regarding the relationship between demographic variables (age, sex, SES, and smoking status) and both periodontal status (as measured by mean CPI score) and ECPEM (malnutrition versus no malnutrition). In addition, we examined the two groups of participants based on ECPEM regarding mean CPI, and tested for an independent association between the periodontal disease severity (as measured by mean CPI score), and the primary exposure of interest ECPEM, controlling for differences in age, sex, SES, and smoking using least squares multivariate linear regression.


The distribution of the 96 study participants by age group, sex, and baseline SES level is shown in Table 1. The mean age of the group was 14.8 years (SD = 1.7 years), and males constituted slightly more than half (56.3%) of the sample. In bivariate analyses of the interrelationship of these demographic factors, we found that smoking was related to increased age; in fact, by the age of 16–19, half of the children reported smoking (χ2 = 15.4, d.f. = 4, P = 0.004). No other important relationships were found between any other of the demographic variables.

Table 1
Demographic characteristics of the study participants.

Demographic characteristics of the study participants by ECPEM are shown in Table 2. Overall, 49.5% of those in our sample were malnourished, which did not differ substantially from the proportion of children who were malnourished in the core study, where ECPEM was recorded in approximately 45.0% of the participants. ECPEM did not vary by sex or age. There was an inverse relationship between SES and ECPEM – those in the lowest SES category (‘poorest’) were more than four times more likely as those who were merely categorized as ‘poor’ to have had ECPEM (OR = 4.2; 95% CI 1.2, 14.3); those in the middle (‘poorer’) category were twice as likely to have had ECPEM; however, the difference between the ‘poorer’ adolescents and those classified as merely ‘poor’ did not reach statistical significance (OR 2.0; 95% CI 0.4, 10.9).

Table 2
Early childhood protein-energy malnutrition (ECPEM) Haitian adolescents.

In bivariate analysis (Table 2), we found that although participants who were malnourished during the first 5 years of life had a higher mean CPI than those who were not malnourished, the difference in mean CPI was not statistically significant (1.63 vs 1.42, Student’s t-test P = 0.10). Although mean CPI scores were reduced when we substituted the alternate CPI score which substituted buccal and lingual surfaces for the mesial surfaces that might have been subject to deepening because of erupting premolar teeth, we found no statistical difference by ECPEM status (1.26 vs 1.09, Student’s t-test P = 0.13).

Results from multiple linear regression analysis, controlling for the potential confounding variables of age, sex, SES, and smoking, however, showed that children who were malnourished in early childhood had worse periodontal status, as measured by CPI, than those who were not malnourished (Table 3). In our final model, which was adjusted for sex and smoking, we found that ECPEM was significantly and directly related to mean CPI. Based on our model, in a non-smoking female without ECPEM exposure, the predicted mean CPI was 0.44, whereas for a smoking male with ECPEM exposure, the predicted mean CPI was 0.87. Calculating mean CPI based on buccal and lingual scores for first molars in those participants under the age of 15 did not meaningfully affect the findings.

Table 3
Linear regression modelling results, ECPEM (independent variable) on mean CPI score (dependent variable).


Over a third of the world’s children are affected by PEM17, a syndrome of inadequate intakes of protein, energy, and micronutrients. The relationship between ECPEM and acute infections is a complex one, as ECPEM adversely affects the immune system leading to an increased risk of infection5,13,18, while, conversely, infection may cause nutritional imbalances leading to PEM5,13,19. The appreciation of this complex relationship has led to the contemporary view of ECPEM being a syndrome, which includes protein, energy and micronutrient intake deficiencies as well as frequent infectious disease episodes17.

It has been hypothesized that ECPEM can have effects that reach beyond early childhood, for those who survive malnutrition. Specifically, ECPEM is thought to affect the developing immune system and therefore may alter an individual’s ability to respond to infection long after the PEM has itself subsided12. This retrospective cohort study of the effects of PEM during the first 5 years of life on periodontal status is unique as reliable anthropomorphic assessments were reliably recorded in a cohort highly exposed ECPEM. Haiti, the poorest country in the western hemisphere, has one of the highest rates of childhood malnutrition in the world: between 40% and 75% of children are reported to have some level of malnutrition20. In our study sample, half of the children had experienced malnutrition, either regularly or intermittently, during their first 5 years of life.

We found that those children who had been malnourished during the first 5 years of life had higher CPI scores than those who had not been malnourished. The effect of EC-PEM on mean CPI score in our model was roughly as great as the effect of smoking, an established risk factor for periodontitis. It is important to note that, in the light of the delayed eruption of the permanent teeth that has been identified in these Haitian children who were exposed to ECPEM, the effect of ECPEM on the periodontal condition of the permanent dentition may, in fact, be underestimated, as the time these teeth were at risk for developing periodontal disease was reduced1.

It is also notable that although determination of the prevalence of periodontal disease level was not the primary purpose of this study, we found that roughly half of all examined children exhibited CPI scores of 3 or greater in at least one sextant. Although we chose to record CPI scores of 3 or greater in those under the age of 15, the mean CPI score did not change appreciably when we substituted sites that would be less likely to be subject to deepening due to erupting teeth. Indeed, CPI scores of 3, which indicate pocketing of 4–5 mm, have not been previously reported in such a young population. Of the two previous studies which investigated dental disease levels in Haitian children, one did not include measures of periodontal destruction21, whereas the other22, conducted over 20 years prior to the present investigation, measured periodontal disease levels using the Ramfjord’s Periodontal Disease Index, which makes comparing the two studies difficult. Nevertheless, this prior investigation identified low levels of disease in Haitian children aged 12–15, with a mean PDI of only 0.59 on a scale of 0, which represents no disease, to a score of 6, which represents a pocket measuring >6.0 mm. Periodontal surveys conducted on children elsewhere in the Caribbean (see Table 4) also show lower levels of periodontal disease than what was found in this investigation2328. Because the purpose of our study was to investigate the relationship of ECPEM and the subsequent development of periodontal disease, and not to quantify levels of periodontal disease, our sampling measures varied from the other studies that have been conducted. In addition, although the CPI is often used to quantify periodontal disease levels, its properties do not make it an ideal periodontal index to establish the presence of periodontitis. But, given the high proportion of adolescents in our study with periodontal pocketing, clearly, further investigation regarding levels of periodontal disease in these Haitian children is warranted.

Table 4
Periodontal status of Caribbean children.

Although we found that SES was related to ECPEM, we did not find that SES was related to CPI score, and in fact in our multivariate analysis which investigated the relationship between ECPEM and CPI score, SES was dropped from the final reduced model, while smoking status and sex remained significant predictors of CPI score. We hypothesize that ECPEM is a likely a good reflection of SES in this population, and that ECPEM is a stronger predictor of CPI than SES. It is perhaps more likely that malnutrition would biologically affect the periodontium via effects on the developing immune system than would SES factors, which may or may not be related to oral habits and oral care in this population.

To our knowledge, this is the first study to investigate the relationship between ECPEM and the subsequent deterioration of periodontal status in the permanent dentition. Although effects of specific nutritional deficiencies on the periodontium have been reported previously810, the mechanism by which PEM in early childhood is likely to be quite different than that hypothesized to exist when the two conditions, nutrient deficiency, and periodontal disease or gingivitis occur simultaneously. Because ECPEM is likely to affect the developing immune system, a person’s ability to respond to colonization with periodontal pathogens may be adversely affected permanently.

What this paper adds

  • In this study sample, a substantial proportion of rural Haitian children aged 12–18 had manifestations of periodontal disease, as measured by CPI. Roughly half of all examined children exhibited CPI scores of 3 or greater in at least one sextant.
  • Early childhood protein-energy malnutrition was related to increased CPI score in this cohort of children.

Why this paper is important to paediatric dentists

  • Results of this study indicate that early childhood protein-energy malnutrition may predispose to periodontitis in the permanent dentition.


We acknowledge the contribution to this study of Rudolph St. Jean, Project Coordinator; Dr Jerry Lowney, founder and Executive Director of the Haitian Health Foundation; Dr Ernst Joseph, Dean of the University of Haiti Dental School; and the key members of field team: Ms. Ruthie Barrieau, an MS in Clinical Research graduate student at NYU, and Dr. Butler Brice, a dentist in the Jeremie community. In addition, we thank Dr Khalid Almas, of the University of Connecticut Health Center, for his assistance in standardization to the WHO CPI. This study was supported by the NIH NIDCR, grants R01 DE014708-01A2, T32 DE007255 and by the New York University College of Dentistry.


1. Psoter WJ, Gebrian B, Prophete S, Reid BC, Katz RV. Effect of early childhood malnutrition on tooth eruption patterns in Haitian adolescents. Community Dent Oral Epidemiol. 2007;36:179–189. [PubMed]
2. Sawyer DR, Nwoku AL. Malnutrition and the oral health of children in Ogbomosho, Nigeria. ASDC J Dent Child. 1985;52:141–145. [PubMed]
3. Cleaton-Jones P, Richardson BD, Granath L, et al. Nutritional status and dental caries in a large sample of 4- and 5-year-old South African children. S Afr Med J. 2000;90:631–635. [PubMed]
4. Alvarez JO, Eguren JC, Caceda J, Navia JM. The effect of nutritional status on the age distribution of dental caries in the primary teeth. J Dent Res. 1990;69:1564–1566. [PubMed]
5. Enwonwu CO, Falkler WA, Jr, Idigbe EO, et al. Pathogenesis of cancrum oris (noma): confounding interactions of malnutrition with infection. Am J Trop Med Hyg. 1999;60:223–232. [PubMed]
6. Enwonwu CO, Phillips RS, Ferrell CD. Temporal relationship between the occurrence of fresh noma and the timing of linear growth retardation in Nigerian children. Trop Med Int Health. 2005;10:65–73. [PubMed]
7. Jimenez LM, Duque FL, Baer PN, Jimenez SB. Necrotizing ulcerative periodontal diseases in children and young adults in Medellin, Colombia, 1965–2000. J Int Acad Periodontol. 2005;7:55–63. [PubMed]
8. Nishida M, Grossi SG, Dunford RG, Ho AW, Trevisan M, Genco RJ. Calcium and the risk for periodontal disease. J Periodontol. 2000;71:1057–1066. [PubMed]
9. Dietrich T, Joshipura KJ, Dawson-Hughes B, Bischoff-Ferrari HA. Association between serum concentrations of 25-hydroxyvitamin D3 and periodontal disease in the US population. Am J Clin Nutr. 2004;80:108–113. [PubMed]
10. Gomez-Carrasco JA, Lopez-Herce Cid J, Bernabe de Frutos C, Ripalda-Crespo MJ, Garcia de Frias JE. Scurvy in adolescence. J Pediatr Gastroenterol Nutr. 1994;19:118–120. [PubMed]
11. Shils ME, Olson JA, Moshe S. Modern Nutrition in Health and Disease. Philadelphia: Lippincott, Williams & Wilkins; 1999.
12. Cunningham-Rundles S, McNeeley DF, Moon A. Mechanisms of nutrient modulation of the immune response. J Allergy Clin Immunol. 2005;115:1119–1128. [PubMed]
13. Scrimshaw NS, SanGiovanni JP. Synergism of nutrition, infection, and immunity: an overview. Am J Clin Nutr. 1997;66:464S–477S. [PubMed]
14. WHO. Oral Health Surveys, Basic Methods. 4. Geneva: World Health Organisation; 1997.
15. Neely AL. Prevalence of juvenile periodontitis in a circumpubertal population. J Clin Periodontol. 1992;19:367–372. [PubMed]
16. Devin RB, Erickson PI. The influence of male caregivers on child health in rural Haiti. Soc Sci Med. 1996;43:479–488. [PubMed]
17. Onis M, Monteiro C, Akre J, Clugston G. The Worldwide Magnitude of Protein-Energy Malnutrition: An Overview from the WHO Global Database on Child Growth. WHO; 2001. [PubMed]
18. Cegielski JP, McMurray DN. The relationship between malnutrition and tuberculosis: evidence from studies in humans and experimental animals. Int J Tuberc Lung Dis. 2004;8:286–298. [PubMed]
19. Schaible UE, Kaufmann SH. Malnutrition and infection: complex mechanisms and global impacts. PLoS Med. 2007;4:e115. [PubMed]
20. Mock N, Bertrand MC. Nutrition in Haiti: An Analysis of Problems and Solutions. Washington: Agency for International Development; 1988.
21. Psoter WJ, Saint Jean LP, Morse DE, Prophte SE, Joseph JRE, Katz RV. Dental caries in twelve- and fifteen-year-olds: results from the Basic Oral Health Survey in Haiti. J Public Health Dent. 2005;65:209–214. [PubMed]
22. Franz FE, Götze W. Oral health survey in Haitian and Hamburg children aged 12–15. Community Dent Oral Epidemiol. 1983;11:302–307. [PubMed]
23. Collins J, Carpio AM, Bobadilla M, et al. Prevalence of clinical attachment loss in adolescents in Santo Domingo, Dominican Republic. J Periodontol. 2005;76:1450–1454. [PubMed]
24. Alonge OK, Narendran S. Periodontal health status of school children in St. Vincent and the Grenadines. Odontostomatol Trop. 1999;22:18–22. [PubMed]
25. Leake JL, Otchere DF, Davey KW, Bedford WR, McIntyre DO. The dental health of 12-year-old children in Dominica: a report of a survey using WHO methods. J Can Dent Assoc. 1990;56:1025–1028. [PubMed]
26. Vignarajah S. Periodontal treatment needs in 12 and 15 to 19-year-old school children in the Caribbean Island of Antigua, 1990. J Periodontal Res. 1994;29:3242–3247. [PubMed]
27. Garcia-Godoy F, Cordero DA, Sanchez CM, Batista J. Periodontal treatment needs in 12–16 years old children from Santo Domingo. Community Dent Oral Epidemiol. 1986;14:250–252. [PubMed]
28. Wertheimer FW, Brewster RH, White CL. Periodontal disease and nutrition in Trinidad. J Periodontol. 1967;38:100–104. [PubMed]