Maternal oral health status and preterm low birth weight at Muhimbili National Hospital, Tanzania: a case-control study

In Tanzania, the prevalence of severe type of periodontal disease is very low and affects only a minority even though poor oral hygiene is a problem of the majority [1–3]. Severe periodontal disease in the mother has recently been reported to be associated with the occurrence of preterm birth, preterm low-birth-weight (PTLBW) and low-birth-weight (LBW) at full term infant delivery in some populations [4–9]. Furthermore, periodontal pathogens such as Tannerella forsythia (previously called Bacteroides forsythus), Porphyromonas gingivalis, Actinobacillus actinomycetemcomitans, and Treponema denticola have been shown to be significantly associated with PTLBW [10–12]. However, contrary findings showing no differences in clinical periodontal status between PTLBW and term normal birth weight outcome have also been reported [11–13].

Potential mechanisms that were put forward to explain the relationship between periodontal disease (PD) and PTLBW infant delivery was that periodontal infection serves as a chronic reservoir of lipopolysaccharide (LPS) which are responsible for the production of interleukin-1beta (IL-1β), prostaglandin E₂ (PGE₂), and tumour necrosis factor-alpha (TFNα) that are in turn associated with preterm parturition and fetotoxicity [10, 14].

In relation to dental caries in Tanzania, with particular interest to advanced carious lesions presenting as carious pulpal exposure are very common and in most cases the treatment readily available is mainly tooth extraction [15]. Generally, in Tanzania, more women compared to men have more decayed and missing teeth [16]. Also, it has been reported that mothers or pregnant women aged 25 to 49 years in Dar-es-Salaam, have lost an average of two teeth and that more than four teeth were found to have caries [17].

As regards to the microbiology of both dental caries and periodontal disease, some of the microorganisms involved in periodontal lesions such as Porphyromonas gingivalis, Fusobacterium, Prevotella, Eubacterium, Peptostreptococcus, and Spirochetes were found in endodontic lesions and some of these are responsible for the production of interleukin 1, tumor necrosis factor and prostaglandin [18, 19]. Therefore it is considered that, in the event of carious pulpal exposure (CPE), the microbes and their toxins within the pulp tissue (endodontic lesions) would most likely produce interleukin 1, tumour necrosis factor and prostaglandin and exert similar mechanisms responsible for preterm birth as it has been explained for periodontal disease.

Low birth weight (LBW) is a known problem worldwide [20] and appears to be much more pronounced in developing countries. In Tanzania, the prevalence of LBW is about 16%–19% [21, 22], whereas among adolescents only, it has recently been reported to be about 48% [23]. From a recent prospective study conducted in antenatal clinics in Mwanza by Watson-Jones and Coworkers, a prevalence of as low as 8% LBW and 12% preterm birth has been reported [24]. In literature, LBW is a multi-factorial problem that incorporates important factors as early pregnancy ≤ 19 years, malaria, anemia, multiple gestations (twins, etc), Edema-proteinuria-hypertension (EPH) ghestosis, infections: urinary tract infections, recurrent bacterial infections, sexually transmitted infections (STI), human immunodeficiency viral (HIV) infection, intestinal parasitic infections, long standing poor nutritional status of the mother, smoking, physically demanding work during pregnancy, and poverty [20, 25–28].

Low birth weight infant delivery has been a major health problem associated with high morbidity and mortality rates [22]. Both, periodontal diseases and PTLBW infant delivery are problems that can be prevented [5, 18, 22]. The association of high prevalence of gingival bleeding, low prevalence of periodontal pockets, as well as a lot of untreated dental caries, particularly carious pulpal exposure with PTLBW among Tanzanians is not known. Bearing in mind the constraints in resources, a retrospective rather than a prospective study was planned. Therefore, the aim of this study was to elucidate the association between oral health status and preterm low birth weight infant deliveries among Tanzanian-African women at Muhimbili National Hospital in Dar-es-Salaam, Tanzania.

The following are the results based on 373 inpatient postpartum mothers. None of the participants had been exposed to oral health education or periodontal therapy before the study.

A number of features inherent in this case-control study deserve some clarification. For convenience, the study was conducted in the maternity block at the MNH, the place where mothers with PTLBW infants can readily be found. Therefore, cases and controls were recruited from the same place. Due to constraints in resources and the set-up of prenatal clinics, it was difficult to recruit pregnant mothers for interviews and clinical examinations at labor wards before delivery, particularly when considering comfort. For this reason it was regarded as much more appropriate to recruit inpatient post-partum mothers. The study participants were interviewed on the reasons for their admission to the hospital but the information was not used to categorize them into cases and controls. Majority of them had low level of education and did not know exactly the diagnosis of their problems and it is more often taken as a tradition to keep those fine details with the medical practitioners. Therefore, patient's medical records were used to get information on weight of the infants and maternal gestation age at delivery. The criterion used to categorize low birth weight and preterm birth is the one that has been in wide use as recommended by WHO and has previously been applied in Tanzanian setting, therefore, data can be compared broadly [20, 37].

The occurrence of teenage pregnancies was substantial and unplanned home deliveries were few both of which are consistent with earlier findings [38]. In the univariate as well as logistic regression analysis, un-married mothers were more likely to deliver PTLBW compared to those who were married and the risk was more pronounced to about one and a half times in the primiparous mothers. The reason for this finding is not yet established but it is speculated that unmarried mothers might have stress and experience long standing poor nutritional status due to wide spread poverty, factors known to contribute to low birth weight [20, 27, 34, 39]. The effect of hypertension, which appeared to be a significant risk factor for PTLBW among all studied mothers (primi- and multiparous together), was not significant when primiparous mothers were considered separately. It is thought that age might be a contributing factor. When mothers that had PTLBW were compared to TNBW mothers, the PTLBW had a lower mean age, which is in agreement with other findings from elsewhere for example in USA [6]. In this study teenage mothers were more at risk for PTLBW than mothers that were 20 years of age or more and this also is in agreement with what has been found elsewhere [34].

The prevalence of periodontal conditions in this study population represents the natural history of periodontal disease because the participants had no history of being exposed to either oral health education or periodontal therapy. Despite, the relatively high occurrence of microbial plaque, dental calculus and gingival bleeding on probing, the prevalence of severe periodontal disease was low and corroborates with other findings in Tanzania, and neighboring country, Kenya [1–3, 40, 41]. Moreover, majority of the populations in Tanzania and Kenya retained most of their dentition in a functional state even up to the age of 65 years signifying that severe destruction of periodontal tissue is minimal [42, 43].

The time period that had elapsed between parturition and periodontal examination was close to the previously reported time of hospital stay (one to thirty eight days and a mean of about eleven days) for the LBW infants [22]. However, this elapsed time possibly might have resulted to an underestimation of the gingival bleeding scores due to gradual resolution of gingival inflammation, a phenomenon that is common within two to three months after delivery [44, 45]. The finding that PTLBW had significantly higher mean number of sites with gingival bleeding on probing as compared to the TNBW is in agreement to what was found among Caucasian mothers in Brazil [8].

Although periodontal disease profile differ from one population to another, it had earlier been reported that the differences do not conform to the traditional generalization that African and Asian populations suffered more severe periodontal disease than other populations [36, 46]. The occurrence of severe periodontal disease (6+mm) among mothers at MNH was low (about three percent) compared to ten percent among Saudi mothers whereby periodontal disease accounted for four times increased risk for PTLBW [7]. This suggests that there might be some racial and regional differences in the prevalence and risk factors for PTLBW infant deliveries. The indirect approach used to generate information on loss of attachment, "ELA", in this study did not show any significant difference between preterm and term deliveries, and the findings are contrary to what has been reported in the study on "oral conditions and pregnancy (OCAP)" in USA that used a direct approach to assess loss of attachment [6].

In this study, the proportion of participants exposed to ETS was substantial while those exposed to active or direct tobacco smoke was negligible. Information on smoking and alcohol consumption during pregnancy among Tanzanian women is scarce. Exposure to ETS in this study did not appear to be a significant risk factor for PTLBW and is in line with Steyn and Coworker's findings [47]. However, it disagrees with the findings of Windham and Coworkers whereby the exposure time to ETS was set at ≥ 7 hours/day in non-smokers [48]. Since in our study there were extremely few active tobacco smokers and the difference between cases and controls was not significant, this does not contradict in any way other reports that active heavy tobacco smokers (ten or more cigarettes/day) are at higher risk for PTLBW [20]. The intensity of tobacco smoking during pregnancy among Tanzanian-African women has been stated to be two cigarettes per day (unpublished).

It was interesting to note that despite the good and accessible health facilities starting from dispensaries, health centres, district hospitals as well as regional and referral hospitals in urban areas and in particular Dar-es-Salaam, most of the participants, both "cases" and "controls", started prenatal clinic late. The reason for this behavior is not known, though it was observed that late attendees of prenatal care had higher proportion of individuals who were primiparous, un-married and have low level of education. Ignorance, lack of experience and fear among those who were not married might have contributed to the late reporting for prenatal care. Such high rate of late reporting for prenatal care has also been reported among other maternal populations in Mwanza, Tanzania [24].

The findings that both smoking as well as late prenatal care did not appear to be significant risk factors for PTLBW in this study population are in agreement with other reports by Moliterno and coworkers in Brazil [9] as well as Buduneli and collaborators in Turkey [49]. In Brazil and Turkey studies the prevalence of destructive chronic periodontal disease was low and therefore comparable to the one reported in our present study. However, lack of significance for late prenatal care versus occurrence of PTLBW is contrary to what has been reported in relation to either LBW or PTLBW [5, 20, 36] and the possible reason might be attributed to the socioeconomic and nutritional status [23] of the study participants and quality of prenatal care, factors that were not within the scope of the present study. Information based on pregnant mothers attending prenatal care in three reproductive clinics in Dar-es-Salaam showed that active smoking in pregnancy was negligible (0.3%) but ETS was substantial (43%) and that alcohol consumption was 35% (unpublished). None of the pregnant women did take alcohol on daily basis. Mostly they used alcohol on weekends and the amount was 1–2 drinks per week. It is speculated that the study participants at MNH had similar pattern of alcohol consumption, because most of them if not all were also from Dar-es-Salaam. Low or moderate consumption of alcohol on weekends in some populations elsewhere has not been associated with an increased risk for low birth weight [50]. The amount of alcohol consumption that has unfavorable outcome on pregnancy, either preterm birth or small-for-gestation-age (SGA) has been reported to be 3 or more drinks per day [51, 52].

From previously studied population in Tanzania, it has been reported that in about two thirds of the mothers that had low birth weight infants, no associated maternal factors could be detected [28]. After stratification of the potential confounding factors [33], some factors had higher relative risk where confounders were "present" compared to where they were "absent", but none of them appeared to be significant risk factors for PTLBW. This does not mean at all that there were no clinical significance in the factors studied, but rather it underscores need for a more detailed study that will include among others, wider scope of PTLBW potential risk factors, intensity and duration of the exposure.

In relation to ethnicity, the present study population exclusively involved indigenous Tanzanian-African women. In some reports where periodontal disease has been shown to be one of the significant risk factors for PTLBW, majority of the participants were Blacks/African-Americans [4, 6]. However, based on clinical findings from the present study population, there was no evidence to support the notion that maternal periodontal disease is a significant risk factor for PTLBW. This corroborates with other findings elsewhere [53, 54], including women of low socioeconomic status [49], and those who never smoked tobacco [55]. The possible explanation for this might be that different populations may be subject to different risk factors for the development of a specific pathology. Furthermore, some important periodontal pathogens have significantly been associated with PTLBW [11, 12, 56]. Since this was beyond the scope of the present study, it remains a subject for further research in Tanzania, preferably using most modern techniques such as checkerboard DNA-DNA hybridization and polymerase chain-reaction (PCR) techniques [10, 11, 19, 49].

The level of dental caries, expressed as independent components as "tooth-decay (D), missing (M), filling (F), and in a cumulative index as DMFT, and carious pulpal exposed lesions, all appeared to be not significant risk factors for the occurrence of PTLBW. The findings are in agreement with what has been reported from Thailand [5] and Finland [54]. However, lack of evidence in the present study that poor periodontal or dentition status may be a significant risk factor for PTLBW, does not preclude in any way the importance of pregnant mothers to have routine dental check-ups and management as recommended by other authorities because this will arrest the progression of the disease [13, 53, 57]. However, recently also, it has been reported in large randomized trials that treatment of periodontal disease in pregnant women did not change the outcome rates of preterm birth or low birth weight [58]. Even in occasion where the treatment was interrupted, it did not lead to an increased risk for LBW infant when compared to women with no history of periodontal care [59].

Among the limitations of the present study worthy of commenting are: One; the human immunodeficiency Virus (HIV) and helminthic infections that have recently been reported to be associated with higher risk of low birth weight [25, 27], and nutritional status of the mother were not assessed because they were not within the scope of the present study and therefore their impact on this study is not known. Two; the diagnosis of carious pulpal exposure was based on clinical presentation alone and was not confirmed with radiological examination nor was there any microbiological investigation for the presence of periodontopathic bacteria in the involved teeth and hence this might have threatened the validity of this measurement. Three; the study was done at the maternity block in a government national hospital situated in Dar-es-Salaam city and therefore the results cannot be generalized to include for example rural population that has been reported to have a slightly worse periodontal health compared to urban residents [3].

It is concluded that, on the basis of these clinical findings on periodontal conditions and dentition status, there is no evidence to support the notion that periodontal disease or carious pulpal exposure are significant risk factors for preterm low-birth-weight infant delivery among Tanzanian-African mothers at MNH. Significant risk factors for PTLBW were young age, hypertension, and being un-married. Further research that is prospective in nature and that will incorporate periodontal pathogens and mediators of inflammation in a larger representative sample of pregnant women is needed.