Does an Apple a Day Keeps the Doctor Away?

Does an Apple a Day Keeps the Doctor Away? Depends on the type of doctor.

Common Sense Evaluated

It is common knowledge that a healthy diet prevents all sorts of diseases.  But often common knowledge is based on assumptions rather evidence.  It might be a bit hard to look at all of nutrition in a ~ 1 page review today but I thought it would be great to look at all of the studies put together. The July 2014 fruit and veg vs all cause mortality meta-analysis from the BMJ

 The study

The authors made meta analysis of all of the studies that they could find looking at fruit and vegetable intake and mortality.  They found 25 relevant studies with a combined 833,234 patients.  All of the relevant studies were prospective cohort studies. In the period followed up (which varied from 4 to 26 years) there were 56,000 deaths.  They did statistical testing on the pooled results of the studies and found that most of the studies were pretty similar (though one study was heterogeneous but analysis with and without this study, the results were the same). 

Study check list:

reputable journal                                                    - yep

big cohort                                                              - yep

fancy maths that I dont quite understand                 - yep

not much variation between all the smaller studies   - yep

as you can see overall a pretty well run study in my simple check list

The results

The study found a significant reduction in all cause mortality and cardiovascular mortality from both fruit and vegetable intake.  It found that there is a mortality hazard ratio of 0.95% for every serve per day in fruit or veg per day.  this was statistically significant (0.001).

For fruit there was a 0.94% hazard ration for each serve per day and for vegetables there was a 0.95% hazard ration per serve per day (all cause mortality).  So it appears that both fruit and vegetables are roughly equally beneficial in all cause mortality benefits.

Cardiovascular risk was assessed in some of the studies and it roughly found that for each serve of fruit or vegetable there was a significant reduction in cardiovascular risk of 4% per serve.

Cancer risk was assessed by two of the studies and it did not show significant reduction in with increased intake of fruit and veg.  It showed a hazard ratio of 97% with each serve but it didn't have a significant confidence interval (0.9-1.03).

A final quite important point about the studies results is that it found a positive benefit for each additional serve of fruit and veg up to 5 serves a day, after that point there was no additional benefit.

Interpretation

The biggest weakness with a study like this is that it doesn't show causation rather it shows association.  The study has shown an association between increased fruit and vegetable consumption and decreased all cause mortality.  This suggests that there is a link between mortality and eating fruit and veg.  And most of the studies did some work to remove some of the confounding factors.  

But no matter how hard they worked to get rid of confounding factors it is still a leap to say 'if I eat more fruit and veg then I will live longer'.  Rather from the study all that we could say is if you observed two people who had different diets in terms of fruit and veg content where they were similar in terms of all of the controlled confounding factors the person who ate more fruit and veg is likely to live longer.

This study also used data in which patients self report fruit and veg intake.  The diets potentially could significantly change through the study, they could have had significant differences in diet prior to the studies and finally there is likely to be significant miss reporting of diet.  Though reporting miss reporting of diet is likely to push the results toward no significant results.

Soo ummmm, what's the wrap up

This was a good study that has very significant results - that fruit and veg is likely to be good for us.  It also showed that perhaps eating fruit and veg will mean in the future you are less likely to be seeing a cardiologist but probably just as likely to be seeing and oncologist.  But we have to be careful with interpreting these results because they are an association rather than causation.

An apple a day keeps the cardiologist away.

Good Luck and Have Fun

Back Pain and Paracetamol

Back Pain and Paracetamol

Medicine in the News

There has been a lot of news and talk about a recent article that 'shows': "Paracetamol does not help lower back pain" - The Guardian 23/7/14
"Paracetamol may not be effective for back pain" - ABC 24/7/14
"Paracetamol doesn't work for back pain" - SMH 24/7/14
So what does the study show and what does other evidence show?

By Paracetamol_acetaminophen_500_mg_pills.jpg: Michelle Tribe from Ottawa, Canada derivative work: Anrie (Paracetamol_acetaminophen_500_mg_pills.jpg) [CC-BY-3.0 (http://creativecommons.org/licenses/by/3.0)], via Wikimedia Commons

The study

On the 24/7/14 The Lancet  published the article "Efficacy of paracetamol for acute low-back pain: a double-blind, randomised controlled trial"  and called themselves the PACE study (Paracetamol for Low-Back Pain Study).  

The study was a RCT which looked at paracetamol in the treatment of acute lower back pain without neurological symptoms.
The study randomised 1652 patients to one of three groups:
1. regular paracetamol 665mg x2 TDS (3990mg daily)
2. PRN paracetamol 500mg x 2 QID PRN (4000mg daily max)
3. Placebo

The study was run from multiple GP practices through Sydney Australia from 2009 to 2012. Follow up was conducted over three months from onset of treatment. It had good patient retention.  It was well randomised.

It had a neat way of blinding the study which I thought would be good to explain.  It gave each patient a box, containing regular tablets and PRN tablets.  Each groups tablets are below:

• In the regular paracetamol group 
• 2 x 665mg paracetamol TDS
• 2 x placebo PRN QID
• In the PRN paracetamol group
• 2 x placebo TDS
• 2 x 500mg paracetamol PRN QID
• In the Placebo group
• 2 x placebo TDS
• 2 x placebo PRN QID

 Patients were told to take both the regular and PRN tablets regardless of their group.

The Results

There was no difference between all three groups.

That's right, no difference.

'Maybe they didn't measure the right stuff?'

One thing that the news did report was the 'primary outcome' of the study: time till pain free.  Before reading the article I questioned whether this was a good outcome.  Let's face it, paracetamol is not a disease modifying agent.  It wouldn't make sense for paracetamol to stop a nociceptive stimulus .  I would only expect the amount of pain to be less for the duration of the study and maybe not for the time to pain free to be effected.  So I wondered to myself as I read and heard the news if it was simply a case of measuring the wrong thing.

But:
There was no difference in pain score either.

The main goal of the study was to see time to when the pain stopped, but it also looked at pain levels, amount of other therapy used, sleep, work, disability and range of other factors.

In all three groups, regular paracetamol, PRN paracetamol, and regular placebo with PRN placebo there was no significant finding in any of the measured outcomes.  One outcome that I loved reading was the question - 'which of your tablets were real and which were placebo.

  Patient belief \ Treatment group     regular                 PRN                 placebo   
Neither contained paracetamol      113/498 (23%)     137/494 (28%)     156/505 (31%)

Regular contained paracetamol       261/498 (52%)     224/494 (45%)     237/505 (47%)


As-needed contained paracetamol  124/498 (25%)      133/494 (27%)     112/505 (22%)

As you can see it seemed pretty random if patients were able to correctly identify if they had placebo in both, regular or PRN paracetamol.  I thought this was a funny thing to measure but it does show that the patients were unaware of which tablet was active or placebo and that a lot of patients must have falsely attributed a positive or negative effect to a placebo tablet.

Interpretation

This study is pretty convincing that paracetamol is not an effective treatment for simple acute lower back pain.

The study mentions one of its own limits is that a lot of the patients had pretty poor compliance with treatment.  In the first week there was good compliance, but towards the end of the study (like towards week 12) patients were on average taking around 1 tablet a day.  I think that this is a really interesting limit to talk about because it says that 'theoretically there might be an effect of paracetamol if people take it regularly for a long time but even in a well controlled study where patients were reinforced to take the paracetamol most of the didn't.'  I guess the challenge that this raises is that from the study we can't say that strict regular paracetamol will or will not help.  But what we can say from the study is that if we recommend regular paracetamol or if we don't it probably won't have an effect.  

The other main thing to note about interpretation of the study is that this study only looks at one area.  This study doesn't show that paracetamol is not helpful across the board - other studies have shown it to be effective at treat pain in other settings.  Rather this study shows only that it is not very useful in acute lower back pain.

TLDR:

A well run study that looked at paracetamol vs placebo in acute lower back pain found no difference.

So the questions for us now are:

Should we use it anyway?

Did the study miss the effect? 

Is paracetamol just a placebo anyway?

Are you going to read the study?

Good luck and Have Fun!

Laurence out