Emerging Issues in Dentistry: Caring for Patients in the Absence of Evidence

Description of System 

Treating patients in the absence of evidence requires working with and communicating uncertainty without compromising care. This process (Figure 2) is a slight modification of common models.5 The following patient scenario demonstrates how I use the process.

View Large Image Download to PowerPoint

Figure 2. Evidence-based process.

Scenario: A 28 year-old female patient presents to our office with the chief complaint of “bleeding gums.” The patient is early in the first trimester of pregnancy. This is her second pregnancy after suffering a midterm miscarriage about 2 years ago. The patient is taking prenatal vitamins and crystallized ginger for nausea associated with morning sickness. Oral examination reveals moderate periodontal disease (PD) with generalized gingival erythema, isolated 4- to 6-mm pockets, and bleeding on gentle probing. Her dental history indicates sporadic care. The patient has confided to the dental assistant that after reading an online article she began to wonder if she has “gum disease,” which, according to the article could explain the problems with her last pregnancy. She asks us if she has gum disease and with obvious emotion asks, “Is there anything [she] can do to prevent another miscarriage?”

Problem: The problem is locating, evaluating, and communicating the latest information on the relationship between PD and pregnancy outcome (PO) with the end result that the patient will make an informed decision about care.

If we believe it is at least as important to say the right things to patients, as it is to skillfully perform the correct procedures, then we need the best evidence before we offer an answer. At this point we should have no problem telling the patient we would like to research the medical literature related to her question to get the best, most up-to-date information. Sometimes a quick chair-side computer search is possible, but it may be better to ask the patient to return in a day or two, giving us time to “study the latest research.”

Questions: Although a patient problem often yields only one clinical question, it is just as common to find a number of interconnected clinical questions. In this case there are several questions, but time and focus leads to the following 3. Are patients with periodontal disease more likely to have poor pregnancy outcomes compared with patients with a normal periodontal condition? In patients with PD, does periodontal treatment improve PO? Is treatment of PD safe in pregnant women?

Search: A quick search of PubMed on March 1, 2008, with MeSH terms “periodontal disease” AND “pregnancy outcome” limited to reviews gave 21 references. Reading available abstracts of reviews in the past 5 years provided 6 systematic reviews (SRs) (Table 1). Other articles contained reviews but without methodological elements consistent with SR guidelines.12 Full-text articles of each of the reviews were secured and reviewed.

Table 1. Systematic reviews of periodontal disease/pregnancy outcomes

	Author	Year	Type	Conclusion
	Conde-Agudelo et al6	2008	SR and meta-analysis of observational studies	PD during pregnancy increased risk of preeclampsia compared with women without PD.
	Vergnes and Sixou7	2007	SR and meta-analysis of observational studies	Likely association between PD and preterm low birth weight. Possible benefit from intervention
	Xiong et al8	2007	SR and meta-analysis of observational studies and randomized controlled trials	Insufficient evidence to support periodontal treatment during pregnancy for the purpose of reducing adverse PO.
	Vettore et al9	2006	SR of case-control, cohort and clinical trials	No sound scientific justification to recommend screening of periodontal disease in pregnant women as a means to reduce adverse pregnancy outcomes.
	Khader and Ta'ani10	2005	SR and meta-analysis	PD increases risk for preterm birth and low birth weight. No evidence that treatment of PD will reduce risk of preterm birth.
	Scannapieco et al11	2003	SR of case-control, cohort and randomized controlled trials	PD may be a risk factor for pre- birth weight. Periodontal intervention may reduce adverse pregnancy outcomes.

SR, systematic review; PD, periodontal disease; PO, pregnancy outcome.

Evaluate: Several SRs addressing our first question are available, but the quality and strength of the evidence are mixed. Numerous case-controlled studies and cohort studies exist, but there are few randomized controlled trials as reflected in the SRs. Prospective interventional studies generally provide better evidence, but they are more difficult and may pose ethical dilemmas. Heterogeneity of studies was a problem because of differences in how periodontal disease was measured, the populations included, confounding health factors, and determination of pregnancy outcomes. These make it difficult to combine studies through meta-analysis. There was no overall consensus, although a positive association between PD and unfavorable PO appears to exist. All studies concluded the evidence was weak. Our conclusion was the weak association of PD with PO was not sufficient to prove causation.

Our second question seeking evidence for improving PO prognosis through PD treatment is no clearer. Some studies showed treatment of PD during pregnancy resulted in better outcomes while other studies showed no effect. A weakness in the evidence is the heavy reliance on small interventional studies. Small studies may show improvement with treatment without reaching statistical significance. Nevertheless, in light of the evidence we could not conclude treatment of PD improves PO.

When assessing benefits to treatment we must always consider possible harms of therapy. This leads to our third question: Is treatment of PD safe in pregnant women? The most common initial treatment for moderate PD is scaling and root planing (SRP).13 It has been established that SRP may cause bacteremia.14 Bacteremia involving group B streptococcus is thought to contribute to chorioamnionitis, endometritis, and sepsis, as well as to preterm labor and premature birth.15

Some have postulated that normal functions such as chewing may cause bacteremia in patients with PD, apparently negating the argument of risk from SRP,16 but a recent study found chewing does not cause bacteremia in patients with PD17 leaving a sense that bacteremia from SRP may be a unique risk. Reason tells us SRP may lead to bacteremia and increase the chance of poor PO.

Asking a periodontist and an obstetrician about treatment of PD during pregnancy led to the advice that SRP is probably not a problem, but the domains of tradition and experience are hardly the affirmation EBD seeks. Another query of the evidence using the MeSH terms “pregnancy outcome AND root planing AND harm” returned only 1 study, by Offenbacher et al.18 After eliminating the term harm, 2 additional studies were found (Table 2).

Table 2. Results of second search

	Author	Year	Type	Conclusion
	Offenbacher et al18	2006	Pilot study (N = 67), randomized controlled trial	Treatment was safe. PD improved, PO improved
	Michalowicz et al19	2006	Randomized controlled trial (N = 823)	Treatment improves PD and is safe, PO was unaffected
	Sadatmansouri et al20	2006	Randomized controlled trial (N = 30)	Treatment of PD eliminated preterm, low birth weight infants.

PD, periodontal disease; PO, pregnancy outcome.

All studies concluded the procedure was safe. Two of the studies were small and did not include power calculations. One study by Michalowicz et al19 included 823 women and found periodontal condition improved in the treatment group, but without significant differences in PO between the treatment and delayed-treatment groups. There were no studies showing negative effects of SRP on PO. Even though no SRs have been completed, we concluded the evidence from randomized controlled trials supports the safety of SRP during pregnancy.

Script: When developing our script we need to remember the difference between association and causation. We cannot tell the patient whether PD caused her first miscarriage. Likewise we cannot tell her that periodontal treatment will improve her PO. We can tell our patient that SRP, the most common initial treatment for PD, appears to be safe during pregnancy.

In light of the evidence it would be careless and unethical to hold out periodontal treatment as a preventive solution to the tragedy of miscarriage. Because of the level of uncertainty in this case, the selection of words is important when communicating the evidence to the patient. I find it useful to write a scripted answer to the patient's questions before the return visit. I don't recommend reading the script to the patient, but the script can help organize our thoughts. It may also be prudent to give a written summary of our discussion to the patient to reinforce the oral presentation. The summary, including literature references, should be kept in the patient chart as evidence of what was discussed.

Treatment plan: The patient is able to decide whether to treat her periodontal disease with the knowledge of our current best evidence. This evidence-based decision goes well beyond a perfunctory “informed consent.”

Discussion 

Careful evaluation of our best information sources may yield a frustrating weakness in the current evidence for answering a clinical question. I find this a common result of queries regarding emerging issues in dentistry. Because of practice inertia, it is tempting to fall back on the informational domains of tradition and experience, but these are inadequate for the EBD practitioner. Cautious interpretation of existing evidence helps us avoid misleading the patient and results in an ethical treatment plan.

Conclusions