| Sign In to gain access to subscriptions and/or personal tools. |
The Relationship between Periodontitis and Preterm Low Birthweight
1 Department of Epidemiology and Quantitative Methods in Health, Oswaldo Cruz Foundation (ENSP/FIOCRUZ), Rua Leopoldo Bulhões, 1480-Manguinhos, 8° andar, Rio de Janeiro, CEP 21041-210 RJ, Brazil; Correspondence: * corresponding author, mario{at}ensp.fiocruz.br
There is no consensus about the influence of periodontal disease on preterm low birthweight. The objective was to investigate the relationship between periodontal disease and preterm low birthweight. A case-control study with 542 post partum women aged over 30 yrs was conducted. Three groups of cases were compared with non-preterm and non-low-birthweight control individuals (n = 393): low birthweight (n = 96), preterm (n = 110), and preterm and low birthweight (n = 63). Periodontal clinical parameters and covariates were recorded. Periodontal disease levels were higher in control individuals than in cases. The extent of periodontal disease did not increase risk of preterm low birthweight according to 15 measures of periodontal disease. Mean periodontal pocket depth and frequency of periodontal sites with clinical attachment level 3 mm in preterm low birthweight cases were lower than in control individuals. Periodontal disease was not more severe in women with preterm low birthweight babies.
Key Words: periodontal disease preterm low birthweight case-control risk factors
Preterm low birthweight is an important cause of morbidity and mortality in newborns in Latin America. Despite intensive research on the etiology of preterm low birthweight, in over 50% of clinical cases, the cause remains unknown (Barros et al., 1992). It has been claimed that periodontal infection increases the risk of preterm low birthweight. Early evidence has shown that women delivering preterm low birthweight babies were almost 8 times more likely to have periodontal disease (Offenbacher et al., 1996). The "Focal Infection Theory" proposed by Hunter in 1910 was being resurrected. According to Hunters theory, bacteria and their products from local infections could be disseminated throughout the body and cause diseases in other organs (Billings, 1912). Offenbacher and co-workers study was followed by others. Most claimed a positive relationship (Dasanayake et al., 2001; Jeffcoat et al., 2001; Offenbacher et al., 2001). In contrast, Davenport and coworkers (2002) reported no association, and a systematic review reported limited evidence that periodontitis is associated with increased risk for preterm low birthweight (Madianos et al., 2002). A later systematic review highlighted that most studies on the subject had serious methodological limitations that could lead to spurious conclusions (Vettore et al., 2006). The uncertainties on the relationship between periodontal infection and risks for preterm low birthweight prompted us to carry out a methodologically robust case-control study to test the hypothesis. The objective of this study was to investigate the relationship between clinical parameters of periodontal disease and preterm low birthweight in women over 30 yrs old.
A case-control study was conducted based on primary and secondary data gathered from 4 public maternity hospitals in Rio de Janeiro, Brazil. The maternity units are referral centers for high-risk pregnancies. Three definitions of cases were considered: Low birthweight, Preterm birth, and Preterm birth and low birthweight. All case groups were compared with a control group of non-preterm birth and non-low-birthweight mothers. The study was approved by the Committee of Ethics and Research of the National School of Public Health-Oswaldo Cruz Foundation (protocol no. 78/02). Assuming a 15% periodontal disease prevalence among control individuals and 25% among cases, with 80% power and 5% Type I error probability, a study with a selection ratio case/control of 3:1 required 551 participants (Flores-de-Jacoby et al., 1991).
We conducted a pilot study to adapt and test questionnaires. Examiner calibration for periodontal clinical parameters was performed. Kappa test and Intra-class Correlation Coefficient of agreement for periodontal pocket depth were, respectively, We recruited 542 eligible women for the study, achieving 79% power. Data were obtained from structured interviews, medical records, and periodontal clinical examinations conducted by six examiners. Inclusion and exclusion criteria are provided in Appendix 1. Babies delivered before 37 complete wks of gestation were considered as preterm (PB). Estimation of gestational age was assessed from the last menstrual period (Berg, 1991). When last menstrual period data were missing, we used the Capurro score to estimate gestational age (Capurro et al., 1978). Intraclass Correlation Coeficient between last menstrual period and Capurro score was 0.92. Low-birthweight newborns were infants weighing less than 2500 g at birth. All newborns were weighed immediately after delivery, with calibrated scales. Gestational age and infant weights were obtained from medical records. Periodontal clinical measurements included Visible Plaque Index (Ainamo and Bay, 1975), visible calculus, Bleeding on Probing Index (Ainamo and Bay, 1975), Periodontal Pocket Depth, and Clinical Attachment Level measured at 6 sites per tooth (mesiobuccal, buccal, distobuccal, distolingual, lingual, and mesiolingual) for all teeth, excluding third molars. Periodontal pocket depth and clinical attachment level measures were recorded to the nearest higher millimeter by means of the North Carolina periodontal probe (Hu-Friedy®, Chicago, IL, USA), 15 mm in length and 0.35 mm in diameter. Oral plain mirrors (Hu-Friedy®) and a head light (model 8720, Trilhas & Rumos®, Teresópolis, Brazil) were used to facilitate periodontal examinations. The average time interval between delivery and periodontal examination was approximately 1 day. Covariate data were collected from medical records or through structured interviews. Anthropometric and socio-demographic characteristics included height, Body Mass Index, ethnicity, education level, marital status, and information on work and income. The overall level of housing conditions was assessed based on 14 characteristics of housing. The women were questioned concerning smoking and alcohol intake, before and during pregnancy. CAGE (Ewing, 1984) and T-ACE (Sokol et al., 1989) questionnaires were used to detect alcoholism and risky drinking, respectively. The women answered questions about illicit drug use, physical violence, contraceptive methods, and parental desire for and satisfaction with pregnancy. Physical activities during the gestational period were assessed by an EPAL questionnaire (Vasconcelos and Anjos, 2003). Anxiety and depression were assessed through validated Brazilian versions of the Spielberger Trait Anxiety Inventory (Spielberger et al., 1970) and The Minnesota Multiphasic Personality Inventory (Hathaway and McKinley, 1943). Pregnancy information—including gestational age, baby weight at birth, type of birth, sex of neonate, baby length, and proportion weight/gestational age—was transcribed from medical records. Occurrences of hypertension, pre-eclampsia, hepatitis B, anemia, gestational diabetes, urinary infection, and infections during pregnancy were also recorded. Prenatal care attendance was assessed by the modified Kotelchuck index (Leal et al., 2004). (A detailed description of methods is in APPENDIX 1.)
Statistical Analysis
Periodontal Clinical Parameters
Thirteen different definitions for periodontal disease, based on previously published studies on this subject, were used (Fig. 2
Covariates Covariate variables were computed for each participant and then each group. Comparisons between and among groups were tested by the Chi-square test and Fishers Exact test for variables expressed in proportions, and by the Mann-Whitney test and t test for continuous variables. Internal consistencies for depression and anxiety scales were evaluated by Cronbachs a coefficient.
Of 2561 puerperal women invited, 172 (6.7%) declined to participate in the study. Of the 2389 who agreed, 1847 were excluded for different reasons, resulting in 542 selected participants (Fig. 1
Gestational age at birth and birthweight were significantly lower in all case groups (Table 1
Number of teeth, bleeding on probing, and level of calculus were similar in all pairs of cases and control groups, whereas the Visible Dental Plaque scores were significantly higher in control individuals compared with preterm cases (P = 0.042) (Table 2
Mean periodontal pocket depth and clinical attachment level, mean and frequency of periodontal pocket depth 4 mm, clinical attachment level 3 mm, and periodontal pocket depth 4 mm with clinical attachment level 3 mm were significantly higher in control individuals compared with those in the low-birthweight group (P 0.05) (Table 2 3 mm were lower in women with preterm births. Periodontal pocket depth and clinical attachment level mean and mean and frequency of clinical attachment level 3 mm were significantly lower in women with preterm births and low-birthweight babies (Table 2
Unadjusted risk estimates [odds ratio (OR)] were calculated between periodontal disease and preterm low-birthweight outcomes. Periodontitis was more prevalent in control individuals compared with preterm low-birthweight cases by all 13 methods of defining the extent of periodontal disease (Fig. 2
The hypothesis that periodontal disease is a risk factor for undesirable pregnancy outcomes was rejected. Unlike some studies on the subject, clinical measures relating to periodontal disease were not associated with low birthweight, preterm, and preterm and low birthweight. Furthermore, control individuals had significantly more severe clinical periodontal disease. There is no consensus on a uniform definition of periodontal disease in studies on periodontal disease and preterm low birthweight (Madianos et al., 2002; Vettore et al., 2006). To overcome that, we used all 13 measures of periodontal disease used by workers in this research field. For all 13 measures, woman having preterm low-birthweight babies had less periodontal disease than did control individuals. Therefore, multivariate analyses were not performed. Case definition numbers 14 and 15 facilitated the assessment of a possible dose-response relationship between severity of destructive periodontal disease and pregnancy outcomes. There was a steady decrease in estimates of association between periodontal disease and preterm low birthweight among groups with more destructive periodontal disease. This finding provides strong evidence of no association between periodontal disease and pregnancy outcomes, since the dose-response relationship is an important test in epidemiological studies identifying new causal factors. It is interesting to note that, in 5 of the periodontal disease criteria, periodontal disease was protective in the unadjusted analysis. Although this protective effect might disappear in adjusted analysis, future studies should investigate possible effects of the underlying inflammation-related cascade of events caused by periodontal disease on the placental region. The finding that periodontal clinical parameters in cases were not higher than in control individuals is in accord with results from some previous studies associating periodontal disease and preterm low birthweight (Mitchell-Lewis et al., 2001; Moore et al., 2004). In addition, evidence of no association between periodontal disease and preterm low birthweight in this study was also reported by others (Davenport et al., 2002; Lunardelli and Peres, 2005; Rajapakse et al., 2005; Michalowicz et al., 2006). Heterogeneity in diagnostic criteria for periodontal disease in previous studies is an important source of discrepancies (Vettore et al., 2006). The strength of our study is that a wide range of measures of periodontal disease was used to avoid objections about measuring periodontal disease. Approaches included all methods used previously. In contrast to our findings, several studies found positive associations between periodontal disease and preterm low birthweight (Offenbacher et al., 1996, 2001; Lopez et al., 2002; Radnai et al., 2004; Dörtbudak et al., 2005; Jarjoura et al., 2005; Marin et al., 2005; Moliterno et al., 2005). One possible explanation for their positive findings is poor control for potential confounders. In our study, exclusion selection criteria for participants were relevant confounders, such as chronic diseases like diabetes mellitus and previous periodontal treatment. Exclusion of women with chronic hypertension avoided bias of confounding, since anti-hypertensive medications are strongly related to periodontal status. Women using antibiotics in the week preceding the study were excluded because of effects of antibiotics on periodontal tissues. Antibiotics given in pregnancy are usually for genito-urinary tract infections. This might explain the lack of association between genito-urinary infection and preterm low birthweight in the current study. Classification bias is the most common systematic error compromising validity of case-control studies. In our study, gestational age was estimated through the last menstrual period, the gold standard method for gestational age assessment in epidemiological studies (Berg, 1991). Last menstrual period information was missing in 18% of medical records, and the Capurro score was used to estimate gestational age in such persons, with high agreement of diagnostic reliability. Reliable sources of information for gestational age and weight at birth, obtained from medical records, avoided potential classification bias. In summary, periodontal disease is not a risk factor for low birthweight, preterm, and preterm and low birthweight.
Case definitions
The authors thank Mirella Galvão. This investigation was supported by CNPq, CAPES, FAPERJ Grant E-26/170.421/2003, Hu-friedy do Brasil, and PAHO/WHO.
A supplemental appendix to this article is published electronically only at http://jdr.iadrjournals.org/cgi/content/full/87/1/73/DC1. Received for publication December 14, 2005. Revision received September 13, 2007. Accepted for publication October 12, 2007.
Journal of Dental Research, Vol. 87, No. 1,
73-78 (2008) This article has been cited by other articles:
|
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
3 mm in preterm low birthweight cases were lower than in control individuals. Periodontal disease was not more severe in women with preterm low birthweight babies.
0.05). 



