Shingles Vaccine Protects Against Dementia, But How Strong is the Effect? Here's My Attempt to Decode It

It could actually cut off almost half of all dementia cases in women.

Shingles vaccines, Shingrix and Zostavax. Source: Wikipedia Commons.

Causal evidence is hard to come by.

It usually takes a randomized clinical trial (RCT) to demonstrate cause and effect, especially involving medical interventions. RCTs work because randomization evenly distributes the countless number of variables between individuals into the experimental and control groups, ensuring that the results seen are strictly due to the intervention.

But sometimes, we don’t need RCTs to demonstrate causality.

Sometimes, we can capitalize on existing systems to do so.

This is what three studies from the UK, the US, and Australia have done during the past year. They provide sufficient evidence to show in a causative manner that the shingles vaccine protects against dementia.

But the effect size exactly was unclear in these studies. After a deep dive and thought, here’s my attempt to pinpoint the effect size to infer insights on their real-world impact in the battle against dementia.

Study 1: Eyting et al. (2023), The U.K.

(I wrote about this study in “An Unexpected Ally In Dementia Prevention: Shingles Vaccination.” But I’ll describe it again with added details.)

Eyting et al. took advantage of a policy in Wales, U.K., which limited eligibility for Zostavax (i.e., a live-attenuated shingles vaccine) based on the date of birth. Individuals born before September 2, 1933, were ineligible for the vaccine, while those born on or after that date were eligible.

This setting provided a natural randomization. As Eyting et al. explained, “[Apart from] the probability of ever receiving the herpes zoster vaccine, there is no plausible reason why those born just one week prior to 2 September 1933 should differ systematically from those born one week later.”

Eyting et al. followed over 280,000 participants (71-88 years old) between 2013 and 2021, comparing dementia incidence between vaccine-eligible and -ineligible groups. Results showed that, over 7.4 years, the incidence of dementia dropped by 1.3% in the eligible group from about 15.8% to 14.5%.

But this benefit was mainly seen and only significant among females, probably because dementia and shingles tend to affect females more often than males. Specifically, the absolute risk reduction was 2.9% among females eligible for Zostavax compared to ineligible females (Figure 1). Based on Figure 1’s rough guess (the precise number was not reported), the incidence of dementia dropped from 16% to 13% in vaccine-eligible females.

Figure 1. Effect estimates of being vaccine-eligible on new diagnoses of dementia separated by women (left) and men (right). For women, a statistically significant absolute reduction of 2.9% was seen in the occurrence of dementia. For men, no statistically significant effects were seen. Source: Eyting et al. (2023), medRxiv.

Importantly, Eyting et al. ruled out the possibility of healthy vaccinee bias, a confounding factor that’s almost impossible to control in typical observational studies. This bias occurs because vaccinated individuals tend to be more health-conscious and healthier overall.

Eyting et al. mitigated this bias by showing that the vaccine-eligible group (i) only had reduced dementia occurrence and not other diseases and (ii) did not seek any more vaccines or healthcare utilization than the vaccine-ineligible group (Figures 2-3). If healthy vaccinee bias were present, a risk reduction in other diseases and more vaccine uptake would also be observed. And the absence of this bias indicates that the naturally randomized setting works in distributing health-seeking behavior equally between both groups.

All the evidence presented points towards Zostavax as the only explanation for the statistically significant reduction in dementia incidence among elderly women in this longitudinal study.

Figure 2. No statistically significant effects of being eligible for Zostavax on other leading causes of disability and mortality, namely ischemic heart diseases, chronic obstructive pulmonary disease (COPD), stroke, lower respiratory tract infection, falls, diabetes, lower back pain, and cancer (lung, pancreatic, colorectal, and breast). Source: supplementary data of Eyting et al. (2023), medRxiv.

Figure 3. No statistically significant effects of being eligible for Zostavax on preventive health actions, including flu vaccine update, statin and anti-hypertensive uses, and breast cancer screening participation. Source: supplementary data of Eyting et al. (2023), medRxiv.

All that said, one caveat is that Eyting et al. only calculated the % reduction in dementia incidence in those eligible for Zostavax. But not everyone eligible went to get the vaccine. So, what’s the true effect of Zostavax then?

While Eyting et al. did not calculate this, we can estimate it by multiplying the incident reduction with the vaccine uptake proportion.

This works following the reasoning that the risk reduction was driven solely by those who were vaccinated in the eligible group. And this reasoning is valid because the study has conclusively established that the vaccine is the sole causative reason for the reduced dementia incidence.

Here’s the rough math based on the available info: