Caffeine in clay target shooting: Share 2009 study

Seven elite double-trap shooters, all men, each with an Olympic or world championship appearance behind them, try three conditions in random order and under double-blind conditions: 4 milligrams of caffeine per kilogram of body mass, 2 milligrams per kilogram, and a glucose placebo. In each condition they shoot four rounds of fifty clay targets, up to a maximum of two hundred, under competition conditions on an outdoor Olympic standard range.

The score does not improve with caffeine. The total across the four rounds is 150 ± 17.1 targets on the high dose, 151 ± 10.9 on the low dose and 155 ± 15.0 on the placebo, with no difference (p = 0.406): both caffeine doses sit below the placebo rather than above it. No difference across the four rounds either, neither within a condition nor between conditions, although the first two fall when blood caffeine is at its peak and the last two come four or five hours after ingestion.

The same holds for everything else measured. Reaction time, the first movement after the targets leave the trap house, does not change (the lowest p is 0.062). The tracking times of the two targets, from release to the shot, do not change (lowest p 0.137). Resting heart rate stays at 73 beats per minute in both conditions, mean arterial pressure at 106 against 104 millimetres of mercury, and neither mean nor peak heart rate during the competition moves. The anxiety inventory filled in before each competition shows no differences on any of its three scales, cognitive anxiety, somatic anxiety and self-confidence.

The questionnaire given at the end of the study does, however, record the expected side effects on the caffeine days: headache, anxiety and tremor. The frequencies are not reported. The authors call on them as one of the reasons caffeine brought no advantage, given that tremor of the hand and arm interferes with aiming.

Data collection spread over four weeks, with at least seven days between trials to allow for caffeine washout. The order of the three conditions was assigned at random for each participant, under double-blind conditions. In the twenty-four hours before each session the shooters abstained from caffeine and alcohol in any form and from strenuous exercise; the diet of the day before the first trial was recorded in a food diary and replicated before the later ones, and on test days the food was provided by the experimenters. Self-reported habitual consumption amounted to about six cups of instant coffee a day, with a range from one to twelve.

The caffeine came as pharmaceutical-grade tablets, broken into fragments weighed on a laboratory balance to reach each person's exact dose, and given in thirty millilitres of lemon or orange drink in opaque cups. The placebo was glucose tablets broken in the same way and served in the same vehicle. Ingestion took place sixty minutes before the first round, to coincide with peak plasma caffeine. Body mass was recorded on arrival each day, since the dose is calculated from it, and varied between 0.6 and 2.8 kilograms from one day to another in the same shooter.

Before each round, fifteen minutes of seated rest, then resting heart rate from a heart rate monitor and systolic and diastolic pressure to derive mean arterial pressure; heart rate was also recorded during the competition. Pre-competition anxiety came from the Competitive State Anxiety Inventory-2 before the start of each trial.

The competition followed the international federation's rules: two clay targets released together from an underground trap house at each call, with the angle and height of release known to the shooters and kept the same across every round of the whole study. A round lasted 16.2 minutes on average. The score is the count of shattered targets, one point each.

Three digital cameras with a shutter at one two-hundred-and-fiftieth of a second, at twenty-five frames per second, mounted three metres behind the shooter. From the footage, with software that brought the sequence to fifty frames per second, came reaction time, counted from the moment the targets become visible in the image to the shooter's first movement, and the tracking times, from the same origin to the appearance of smoke at the muzzle for the first and the second shot.

Temperature, wind and humidity were measured before each round and did not differ between days: 11.7 degrees, 72.2 per cent humidity, 2.5 metres per second of wind on average. The analysis uses repeated measures analysis of variance, two conditions by four rounds across all seven shooters for the comparison between the high dose and the placebo, and three by four on the five who also did the low dose; normality checked with the Shapiro and Wilk test, threshold at 0.05. Approval from the local ethics committee, written consent from everyone.

Seven people. It is the limit from which nearly everything else follows, and it shows in the effect sizes next to the p-values: between the high dose and the placebo the score gives p = 0.102 with a partial eta squared of 0.304, which by convention is a large effect. At this sample size the study does not separate the absence of an effect from an effect it lacks the power to see, and there is no power calculation and no confidence interval anywhere. What can be said with some support is the direction, since both caffeine doses sit below the placebo; what cannot be said is the sentence in the conclusions, that 4 milligrams per kilogram provide no benefit, taken as a general statement.

No measurement confirms that the caffeine arrived where it was meant to. There is no blood, saliva or urine sample, and not even a physiological response to stand in for one: resting heart rate is identical in the two conditions, 73 against 73, and mean arterial pressure moves by two millimetres of mercury. Four milligrams per kilogram in men of ninety-two kilograms comes to about three hundred and seventy milligrams of caffeine, a dose that usually shows up in cardiovascular measures. The self-reported habitual intake, about six cups a day with peaks of twelve, and an abstention of only twenty-four hours explain why it might not show: these are heavily habituated consumers, partly in withdrawal on the day of the trial, the placebo day included. The null result therefore has two readings the design does not separate, that caffeine does not help, or that in these people it did next to nothing.

The double blind is stated and never tested. Nobody asked the shooters which drink they thought they had been given, and the side-effects questionnaire, administered once at the end of the study rather than after each trial, reports that headache, anxiety and tremor appeared on the caffeine days. Anyone feeling a tremor knows which arm they are in, so on those days the blind can hardly have held. Of those effects, moreover, it is not known how many people had them or at which dose: the sentence citing them as a possible explanation for the missing benefit remains a hypothesis without numbers.

The crossover is incomplete. All seven did the high dose and the placebo, but only five also did the low dose, and the abstract speaks of seven shooters and three conditions without saying so. The three-way comparison therefore runs on five people.

Then there is the point that matters for the registry: these are not archers, and the task is not the same. In double trap the shooter reacts to two targets leaving together and fires within about a second; in archery a position is held against a resistance and the string is let go at a stationary target. A null effect on the reaction to a clay target says nothing about a ten-second hold, and conversely a tremor that does not spoil a quick shot may well spoil an arrow.

The only data on archers the paper cites are unpublished. The discussion returns six times to a study by Cunningham on elite archers, according to which two milligrams per kilogram would raise the score by 5.7 points over seventy-two arrows without reaching significance, and six milligrams would lower it by 7.7. Those numbers are labelled as unpublished data and cannot be checked anywhere, so they do not enter the registry.

Finally, the study measured a great many things, score, reaction, two tracking times, heart rate at rest and in competition, blood pressure, three anxiety scales, across four rounds, and reports no correction for the number of comparisons. It matters little here, because nothing came out; it would matter if something had.

For an archer this study counts above all for one thing: it is the cleanest design in existence on caffeine and precision shooting, double-blind with a placebo, in a real competition, on athletes of Olympic standard, and it finds nothing. No gain in score, no faster reaction, no shorter tracking, no decline in the closing rounds as the caffeine comes down. If caffeine did what its reputation promises in precision sports, this was a place to see it.

The two things that do emerge, and that an archer will recognise, sit at the edges of the main result. The first is the direction: with caffeine the scores are lower rather than higher, at both doses. The difference is not significant and with seven people it would not become so, but it does not point towards an advantage. The second is the side effects reported on the caffeine days, headache, anxiety and tremor. In clay target shooting a tremor is lost inside a movement lasting a second; in a movement where one holds still against twenty kilograms for ten seconds, it is not.

What this study cannot say is how things would go with a bow, because the task is a different one and because the seven people are clay target shooters. Nor can it close the question even for clay target shooting, since seven participants with a spread of fifteen targets leave room for a good deal. Worth noting too that the participants drank six coffees a day on average and abstained only for the preceding twenty-four hours: someone living with that habit is not the person on whom a single dose is best measured.

The registry keeps this work as the firmest evidence available on a precision task, in a sport other than archery and on seven people, and records the contrary direction alongside the side effects. It draws from it no guidance on what to drink before a competition.