Chemical Exposures: No Dental Dilemma for BPA

Among the many uses of bisphenol A (BPA) is the manufacture of resin-based dental composites and sealants. Recently a team of researchers from the CDC sank their teeth into questions about whether BPA monomer leaching from sealants could be harmful to people. The results of their human study, presented in the March 2006 issue of the Journal of the American Dental Association, suggest that although leaching does occur, sealants are still a safe means of preventing dental cavities. 
 
Low-level exposures to BPA monomer in pregnant rodents, at a level that humans could potentially receive from dental sealants, have been shown to disrupt reproductive development in their fetuses, and concerns have emerged about the possibility of human health effects from dental exposures. Scientific exploration of this question has yielded inconsistent results, says Renee Joskow, first author of the March paper. Much of this is due to limitations in laboratory detection and translation of animal studies to human health effects, as well as insufficiently addressing the parameters of exposure in a clinical dental setting. 
 
The CDC team, led by Joskow (now of the U.S. Public Health Service) and Dana Barr, looked at 14 nonsmokers receiving their first resin-based sealants as part of their routine dental care. Each subject received one of two brands of dental sealant manufactured by two well-established dental equipment and material supply firms. Then their saliva and urine were tested for BPA. 
 
All the patients had BPA in their saliva and urine, even before treatment. For patients receiving Helioseal F sealants, saliva BPA doubled immediately after treatment and returned to baseline within 1 hour. Urine BPA more than tripled 1 hour after treatment and returned to baseline within 24 hours. For patients receiving Delton LC sealants, saliva BPA increased nearly 126 times immediately after application and was still 23 times higher after 1 hour. Urine BPA jumped 10 times 1 hour after treatment and was still elevated 24 hours later. Both levels eventually returned to baseline. 
 
Barr believes the patients’ baseline BPA came from background exposures from environmental sources such as water and food packaging. These, she suggests, could be “a more chronic low-level source of exposure” than dental sealants. Barr adds that in her view, although point-source exposure from dental sealants might approach levels that induce health effects in rodents, “[it] is not the most significant source of exposure in humans.” Moreover, she holds that exposure to BPA from dental sealants, already variable and short-lived in the body, could be easily reduced further by having the patient spit frequently in the hours after application.

The successful management of patients with malignant lymphoma is dependent upon accurate staging and an appropriate treatment regime. Although the chest is a frequent site of involvement (Chabner, 1977;Filly et al., 1976;Kaplan, 1980;Peckham, 1973;Rosenberg, 1961), investigation has relied upon the plain chest X-ray (CXR) and tomography (Tomo) until the introduction of computed tomographic scanning (CT scan).
Prognosis has been found to correlate with the presence of very bulky disease in the chest, and even moderate amounts of intrathoracic lymphoma have been associated with a worse response to treatment than would be expected for other sites of involvement (Lee et al., 1980;Mauch et al., 1978). CT has been found to detect more mediastinal and pulmonary disease (Fong et al., 1982;Muhm et al., 1979;Osbourne et al., 1982;Underwood et al., 1979) in patients with thymoma and bronchial carcinoma, and it has been suggested that it may improve the staging of intrathoracic lymphoma (Ellert & Kreel, 1980). Comparison has not previously been made with conventional radiography, therefore a prospective study was undertaken to determine the role of thoracic CT in the management of patients with lymphoma.

Patients
One hundred consecutive, previously untreated patients with biopsy proven malignant lymphoma were studied over an 18-month period ( Table I).
The histological diagnosis was established in all cases by Dr A.G. Stansfeld and the patients were (a) Chest X-ray: posteroanterior (PA), lateral and penetrated PA views. (b) Tomograms: PA and lateral tomograms of the mediastinum and hila at 1 cm intervals. (c) CT scan: CT scan was performed with contiguous 1.3cm sections from the sternal notch to the diaphragm using an EMI 5005 CT body scanner. Intravenous contrast was given to 14 patients to enhance the major vessels when findings were equivocal by plain scan.
In the second half of the study a further 50 patients were examined by PA, penetrated and lateral CXR and CT scan only. For these patients the CT scan included the lung bases only when abnormalities were present in the upper part of the chest (19/50).
Follow up At the completion of treatment all patients were reexamined with repeat PA, lateral and penetrated CXR. CT scan was repeated in all those in whom it was previously abnormal or equivocal.
All patients have had monthly clinical follow up for 24-42 months from the completion of treatment, with further CXRs and CT scans.

Analysis
At the completion of the study the CXR and tomograms (AKT) and CT scans (IKF, FEW) were reported independently. All cases in which the findings differed were then reviewed together with the X-rays taken following treatment and with the knowledge of the patient's clinical course and response to therapy.
Soft tissue masses in the mediastinal and hilar regions were reported as lymphadenopathy. Parenchymal lymphoma was only diagnosed in the presence of discrete nodules or infiltrates and in the absence of clinical evidence of infection. Histological confirmation of parenchymal lymphomas was achieved in 6 patients. Shadows radiating from the hila were not accepted as evidence of lymphomatous infiltration as they could not be distinguished from vascular or lymphatic congestion. Equivocal findings were recorded as negative and post radiation fibrosis was disregarded.
Statistical analysis was made using the Chi squared test with Yates' correction and Student's t test.

Definitions
True positive abnormalities Those confirmed histologically (15 patients), and those showing resolution on repeat examination following successful lymphoma treatment without antibiotic therapy, or progression in association with other sites of disease upon the failure of treatment. One patient in whom histological confirmation was not obtained for the chest abnormalities died of lymphoma during 'treatment. All the radiological findings in this case were in agreement and were assumed to represent a true positive.
False positive abnormalities Those initially thought to be due to lymphoma, but subsequently shown on clinical and radiological follow up not to represent disease (see Results).
True negative investigations Those which were normal at presentation and remained so at follow up examination with no clinical evidence of disease elsewhere.
False negative investigations Those which were contradicted by other radiological findings which fulfilled the true positive criteria.

Results
Comparison of the results from CXR, tomography and CT scan In all instances CT scan proved to be the most sensitive and CXR the least sensitive method for detecting intrathoracic lymphoma. The results of conventional tomography were intermediate between those of CXR and CT scan but never significantly more sensitive than the CXR. Tomography was therefore stopped after the first 50 patients (Table II). The findings from CXR and CT scan differed most significantly in the mediastinum (P<0.001 for 100 patients (Figure 1; Tables III and IV)). The increased detection by the CT scan was particularly   Table V). CT scan, although more sensitive than CXR, was no more accurate at detecting hilar lymphadenopathy than tomography (Table IV). This was due to the difficulty experienced with all methods in differentiating enlarged nodes from hilar vessels.
False positives in the mediastinum (Tables II and  III) were due to prominent vessels in all four cases as shown by i.v. contrast enhancement of the CT scans (Figure 2). Similarly at the hilum four of the false positives were due to vessels and one, on tomography, to a large anterior mediastinal mass which was wrongly interpreted as involving the hila.
Pulmonary parenchymal disease was detected with greater sensitivity by the CT scan compared with the CXR (P<0.05, Table IV). All 4 patients in whom CT scanning detected occult parenchymal disease had NHL with no other evidence of extranodal spread (Figure 3). Biopsy confirmation was obtained in 2 of these 4 patients.
Effect of CT scanning on staging and treatment Thirty-nine patients had evidence of Stage IV disease prior to CT scanning and thus neither stage nor the treatment was affected by the CT scan.
There was a change in stage in 10/61 patients with Stage I-III disease (16%) due to information solely available from the CT scan. The stage was increased in 9 patients: 4 from II-IV by the detection of parenchymal deposits (2 confirmed on biopsy), 3 from I-IT, and 2 from II-III by the detection of intrathoracic lymphadenopathy. Stage was decreased in one patient from III-I when the CT scan demonstrated that a mediastinal shadow was due to normal vascular structures.   Thirty-two Stage 1-111 patients had either B symptoms or very large intrathoracic masses for which chemotherapy is our current treatment of choice, and treatment was not therefore affected by CT scan results. Treatment was changed in 11 of the remaining 29 patients (38%). Four required chemotherapy for parenchymal deposits and one (downgraded from III to I) received involved field radiotherapy for an inguinal deposit. Six other patients receiving radiotherapy to the involved field for NHL had the field changed by CT scanning (includes one patient in whom the stage was not changed). No change was made in those with localised HD as all received mantle field radiation which adequately covered any additional sites.
CT scanning and remission assessment (Table VI) Complete remission was achieved by CXR criteria in 28/46 patients with intrathoracic lymphoma, although the CT scan returned to normal in only 18. There have been 9 relapses from the 28 with normal CXR (32%) of which 6 were correctly identified by the remission CT scan as having residual disease after treatment. In comparison, only 3/18 (17%) in whom the CT scan became normal have relapsed.
However, a positive CT scan following treatment did not always represent residual disease. Eight of the 28 in whom the CT scan was still positive have not relapsed in 2 years. A post treatment laparotomy was performed in one patient ( Figure  4) with marked residual abnormalities, but only fibrotic and hylanised tissue was found.

Discussion
With the introduction of any new staging technique the two principal questions that must be answered are: how does it compare with current practice and what is the clinical relevance of any additional findings? We have attempted to answer these questions in a series of one hundred consecutive and previously untreated patients as a representative sample of our general lymphoma practice. The addition of CT scanning to the investigation of the chest increased the frequency with which intrathoracic lymphoma could be detected to 46% in newly diagnosed patients, greater than hitherto reported (Chabner et al., 1977;Filly et al., 1976;Kaplan, 1980;Peckham, 1973;Rosenberg, 1961).
CT scan was shown to be the most sensitive technique in the assessment of mediastinal and parenchymal disease particularly in patients with NHL but also in HD. However CT scan displayed only a small advantage over CXR or tomography at the hilum as has been found for other tumours within the chest (Fong et al., 1982;Muhm et al., 1979;Osbourne et al., 1982;Underwood et al., 1979;Baron et al., 1981). The low incidence of pericardial and pleural infiltration reflects the early stage in the disease at which most of these patients had been diagnosed.
The additional information from CT scan changed the stage in 10 patients and the treatment in 11. The relative importance of these numbers, however, becomes more apparent when the size of the group "at risk" is taken into account; i.e. change in stage 10/61 (16%), change in treatment 11/29 (38%). This compares favourably with the yield from other staging procedures such as lymphography or bone marrow biopsy (Chabner et al., 1977;Kaplan, 1980). The value of CT scanning at remission assessment was more difficult to quantify, particularly as relapse was the only positive criterion available. The return of the CT scan to normal was a better prognostic factor than CXR. However, residual abnormalities on CT scan after treatment did not always represent active disease, as has been found at post treatment laparotomy (Sutcliffe et al., 1982).
In conclusion, the greater sensitivity and specificity of the chest CT scan produced changes in staging and treatment which were most relevant to the management of those patients with localised disease or only minor abnormalities by conventional chest radiology. Further follow up will be necessary to assess the relevance of CT scan at remission assessment. The difficulty in providing a CT scanning service for the routine staging of these patients with malignant lymphoma argues for the greater concentration of lymphoma treatment in specialist centres.