Why a headline figure tells you so little
Imagine two articles. One says people with gum disease are “twice as likely” to develop rheumatoid arthritis. The other says the increase is “less than one percentage point”. Both could be describing exactly the same underlying data.
That is not a trick. It is the difference between relative and absolute risk, and it is the single most common way health statistics mislead without technically lying.
Take a purely made-up illustration — these figures are arithmetic to show the mechanism, not findings from any study:
| Way of putting it | The same imaginary data |
|---|---|
| Absolute risk, group A | 4 people in 1,000 |
| Absolute risk, group B | 8 people in 1,000 |
| Relative increase | “double the risk”, “100 % higher” |
| Absolute increase | 4 extra people in 1,000, i.e. 0.4 percentage points |
| Turned around | 992 of 1,000 in group B still unaffected |
Every row is honest. The one that gets into a headline is row three. If you only remember one thing from this article, make it this: ask what the absolute numbers were before you react to a relative one.
Four questions to ask of any risk figure
1. Relative or absolute? Covered above. Relative figures make small differences look dramatic; absolute figures tell you what it means for a person.
2. Risk or odds? Much research reports odds ratios, which are not the same as risk, even though almost every news report treats them as interchangeable. When the condition being studied is rare, the two are close enough not to matter. When it is common, an odds ratio exaggerates the apparent effect. Both appear in this field, and the distinction is rarely explained.
3. Compared with whom? A figure describes a difference between two groups, so the comparison group determines the result. Dental patients compared with the general population, hospital patients compared with healthy volunteers, smokers compared with non-smokers — different comparisons, different numbers, same underlying biology.
4. Measured how? This one is specific to gum disease and it is a genuine problem. There is no single definition of periodontitis used across all studies: some rely on pocket depths, some on attachment loss, some on bone visible on X-rays, some on how many teeth are affected, some on patients simply reporting they have gum problems. When two studies “measure periodontitis” and mean different things, their numbers cannot be sensibly averaged — and yet they often are.
What "adjusted for" can and cannot do
Any decent study adjusts its numbers for factors that could distort the result — typically age, sex and smoking. That adjustment is real statistical work and it improves the estimate.
It does not make the problem disappear. Adjustment relies on how accurately the confounding factor was recorded, and smoking is recorded notoriously imprecisely: current or former, how many years, how many a day, and answered honestly to a researcher. Whatever the adjustment fails to remove is called residual confounding, and in this field it is the standard caveat, because smoking independently raises the risk of both conditions: Smoking, rheumatoid arthritis and your gums.
There is also a limit no adjustment can overcome. Adjusting for a factor tells you nothing about the direction of the arrow. A study can show a strong, well-adjusted association between gum disease and arthritis and still not tell you which came first, or whether something unmeasured produced both. That question is dealt with separately in Can gum disease cause rheumatoid arthritis?.
The numbers that are actually about you
Here is the part with practical value. Study averages describe populations; the following describe you, and most of them are measured in a single appointment.
Pocket depth, in millimetres. A thin graduated probe is walked around each tooth and the depth of the gap between tooth and gum is recorded at several points. A healthy sulcus measures only a few millimetres. Deeper readings mean the attachment has been lost, and that is what defines periodontitis in your specific mouth rather than in a study population — and what determines whether Periodontal treatment (Parodontitis) – service page (EN) is indicated.
Bleeding points, as a percentage. How many of the measured sites bleed when gently probed. This is the most useful single indicator of current inflammation, and unlike almost everything else in this field, it responds within weeks to changes in treatment and home care. It is also the number to watch over time: your own value now versus your value six months ago is a far more informative comparison than any population figure.
Number of teeth, and bone level on the X-ray. These record the cumulative damage — what has already happened, rather than what is happening now.
On the medical side. Your GP (Hausarzt/Hausärztin) tracks general inflammation via CRP and related values (Entzündungswerte). Your rheumatologist tracks joint disease activity with a composite score — in Germany usually DAS28, which combines examined joints, your own assessment and a blood value. These are the numbers that describe your arthritis. Your dentist does not interpret them, and should not.
Why this matters more than it sounds. A study can tell you that a group like you has a somewhat higher probability of something. Your own bleeding score tells you whether you have an active inflammation in your mouth today, and whether what you are doing about it is working. One of those is actionable.
Reading a health headline in thirty seconds
A short checklist that works for this topic and for most others:
1. Does the headline use a relative figure (“twice as likely”, “40 % higher”)? Assume the absolute difference is much smaller until shown otherwise. 2. Is the claim about association or about cause? Look for the verbs: “linked to” and “associated with” are association; “causes”, “leads to”, “triggers” are claims that need much stronger evidence. 3. How many people, over how long? A study of a few dozen patients over three months cannot answer a question about a disease that develops over years. 4. Who is being compared with whom? 5. Is somebody selling something at the end of the page? Bacterial test kits, supplements and “focus removal” programmes are the usual candidates in this particular field.
Applied to gum disease and rheumatoid arthritis, this checklist reliably produces the same verdict: consistent association, plausible mechanism, causation unproven.
Which of your own numbers should prompt an appointment
Because an honest answer does not exist in the form the question expects. Published estimates vary depending on the population, the definition of periodontitis and the adjustments made, and no figure derived from a study group applies to an individual. Quoting one would look more precise and be less true.
Rheumatoid arthritis is diagnosed considerably more often in women, which is established. Severe gum disease is generally more common in men. Those two facts pull in opposite directions, which is one more reason to be sceptical of any single combined figure. Your own examination findings are unaffected by either statistic.
It is a measurement, not a verdict, and it needs the rest of the context: how many sites are that deep, whether they bleed, what the X-ray shows and how it compares with last time. A few slightly deeper sites in a stable mouth is a different situation from the same reading everywhere with bleeding throughout. Ask which of the two you are.
No routine blood test does that. CRP shows inflammation somewhere in the body without localising it. Genetic and bacterial tests exist and have narrow, specific uses decided case by case — none of them replaces measuring your gums.
The uncertainty is about the link to arthritis, not about gum disease itself. That untreated periodontitis damages the bone around teeth and eventually costs teeth is not in doubt at all. The uncertain part is the bonus question; the certain part is reason enough to act.
Numbers in this field deserve respect and scepticism in equal measure. The association between gum disease and rheumatoid arthritis is real enough to be worth studying and too variable to be worth quoting at a patient. The figures that should influence what you do are your own: how deep your pockets are, how much your gums bleed, how that compares with six months ago. Those you can measure, and those you can change.
Zahnarztpraxis Tsypin · Wuppertal. This article was medically reviewed and is provided for information; it does not replace an examination. Updated August 2026.
