[skadamou]

"We found no evidence of increasing risk with a larger area of total tattooed body surface."

Without a dose response, I'm inclined to believe that the increase in lymphoma seen in people with tattoos has more to do with confounding factors than with the ink or the act of getting a needle poked into your skin. I would think that controlling for all confounders in a study like this would be exceptionally difficult.

That said, I'm pretty sure that at least some inks do contain known carcinogens[1]

[1]https://tattoo.iarc.who.int/background/

[gwern]

This is a Swedish study, so what might be possible is using the population registry to contact siblings of the cancer patients to ask about traits like tattooing and then their health data would already be in Swedish system and linkable. This would control for a lot of the relevant confounders.

[tempestn]

Though not those related to people's choice to get tattoos.

[esperent]

Tattoos are so common these days that I'm not sure you can say much about a person just because they have tattoos.

[tempestn]

Any given person, for sure. In aggregate though I bet you'd find correlations.

[robin_reala]

Not sure you could in Sweden. Stats[1] show 21% tattoo prevalance in Sweden compared to 12% in the EU in general.

[1] https://www.ncbi.nlm.nih.gov/pmc/articles/PMC10163470/

[tyre]

they are statistically cooler

[rightbyte]

Which is a strong confounding factor ...

"Sharing traits common in motor cycle gangs increases risk of cancer - new study"

[gwern]

No, it would control for a lot of those factors. Things like impulsivity, sexuality, sensation-seeking etc. All very heritable or family-level. If you ever find any correlations reported for tattoos between family members, it's gonna look like everything else: steeply increasing with relatedness.

[tgv]

Even a correlation with the amount of ink could be a lifestyle confound. I'm pretty sure that the population that has a small tattoo differs from the one with large parts covered. Indeed, it is hard to find a cause.

[timr]

Yes. Also, the survey response rate was the biggest difference between groups (54% vs 47%), which could easily explain the observed differences. The confidence intervals cross 1.0 for nearly all reported IRR values.

For those who don't know how to interpret medical evidence, this study is very weak.

[usgroup]

Those response rates are fairly awful with two groups that are markedly different. Seems very likely that they’d self-select on the face of it especially if they knew what the research question was.

[downrightmike]

Tattoos basically shift your immune system into overdrive https://www.youtube.com/watch?v=nGggU-Cxhv0

[office_drone]

Indeed. It's not the ink content that led to Am J Clin Pathol. 2014;142(1):99-103. saying:

"The mean age of death for tattooed persons was 39 years, compared with 53 years for non-tattooed persons (P = .0001). There was a significant contribution of negative messages in tattoos associated with non-natural death (P = .0088) but not with natural death."

[refulgentis]

I'm not sure "people with negative msgs in tattoos died 14 years earlier" sheds light for me on the TFA.

TFA has a more direct, physical, concern - it starts from a well-known, that tattoo ink ends up in lymph nodes, and it does a statistical analysis showing there's a significant statistical result in lymphoma occurence.

I think people with negative tattoos dying younger reduces the # of people with tattoos who get lymphoma, as they have less ink-in-lymph-nodes years.

[marcosdumay]

It shows the existence of some very strong confounding mechanisms.

[refulgentis]

There's certainly plenty of those! :)

I doubt they intended to communicate something that general, and if they did, I doubt they meant to pick one that would reinforce the conclusion.

[usgroup]

Yeah totally agree. That the size of the tattoo or the number of them not increasing risk makes no sense. Somewhat like claiming whether you smoke a cigarette or 20 a day, the risk is the same. If the latter was true it would more likely indicate that there is some other commonality in that group increasing the risk.

Also the slicing and dicing, “11 more than the index year” and so on, is multiple hypothesis testing on the face of it; I wonder if they adjust for that.