Seven studies, ten comparisons, one pooled answer: isolation and compound exercises built similar total muscle. But a similar total isn't the same shape.
The extension grew the long head 17.5%, the bench press 2.1%. The lateral head flipped: 7.2% for the bench press, 0.6% for the extension.
The argument runs on identity. One side calls themselves minimalists — bench, rows, overhead press, done. No curls, no extensions. Efficient. The other side stacks isolation on top of every compound, convinced their arms look different because of it. Both have been running this debate at squat racks and in comment sections for over a decade.
A research team led by Avery Rosa decided to measure which side was right. They pooled seven studies comparing isolation exercises like curls to compound exercises like bench press, tallied ten direct comparisons, and measured one thing: did one type build more muscle?
The measured gap was 0.07, which the researchers class as trivial. They also found no evidence of appreciable heterogeneity between the studies. That is not the same as every study coming out the same: in one, curls grew the elbow flexors significantly more than dumbbell rows did, 11.06% against 5.16%.
Pooled, the answer was still clear enough for the authors to conclude that the evidence indicates similar whole-muscle growth from the two types. For total muscle size, the minimalist's efficiency argument holds.
But total muscle size isn't the only thing the researchers measured.
The compound-vs-isolation debate had no winner — because both sides were measuring different things. Similar total growth. Different regional architecture.
- Across seven studies and ten comparisons, isolation and compound exercises produced similar total muscle growth, a trivial pooled difference.
- One study went deeper and found that different exercises grew different heads of the same muscle more — the extension the long head, the bench press the lateral head.
- Studies where isolation added extra sets came out much like those with matched volume: the difference was trivial, though the estimate was imprecise.
- Most of the evidence covers upper body muscles. For the hamstrings and calves, no study has made the comparison, and the authors expect compounds to fall short for them.
- The data reframes the debate: compounds handle total size, isolation targets regional emphasis. The choice depends on what you're training for, not which side is 'right.'
Different Exercises, Different Heads
One of those seven studies went deeper. Brandão and colleagues tested 43 men over ten weeks, comparing the bench press to a triceps extension. They didn't just measure how much the triceps grew. They measured which parts grew.
The extension grew the long head 17.5%. The bench press grew it 2.1%. An 8.3-to-1 difference.
The bench press grew the lateral head 7.2%. The extension grew it 0.6%. A 12-to-1 difference, this time favoring the compound.
The totals did not match in this study. Measured by MRI, the whole triceps grew 9.5% with the extension and 4.8% with the bench press, which was not a significant increase. The medial head leaned the same way, 14.0% against 7.3%, a gap the authors call statistically similar. Pooled across all seven studies the totals came out similar; inside this one, each exercise grew different heads more.
The minimalist was measuring total size and seeing no difference. The maximalist was noticing contour and seeing a big one. They were arguing about different things. The question was not only "do curls build more muscle?" It was also "which part of the muscle do you want to grow?"
Brandão measured the triceps. The same regional split appeared in a completely different muscle. An eight-week trial that put each participant's legs on different quad exercises found the outer quad grew 8.7% with leg press but only 2.9% with extension — opposite regions of the same muscle responding to opposite exercises.
The Extra-Sets Defense
One last argument for the maximalist: "I do more total sets because I add isolation on top of compounds. Surely that extra volume builds more total muscle."
The researchers tested it directly. Studies where adding isolation meant extra sets came out much like studies where the volume was matched: splitting them that way showed only a trivial difference (0.14), though the authors note the range around it was relatively wide. The extra sets did not show up as clearly more total growth. Curls add time. Whether they add size on top of compounds, these seven studies could not show.
Where This Applies — and Where It Doesn't
Six of the seven studies measured upper body. Only one included lifters with training experience. No study tested hamstrings. No study tested calves. The similar-growth finding rests on arm muscles and one study of the thigh.
For the hamstrings and calves, the authors expect compounds to fall short: squat training grows the hamstrings negligibly, and since compound leg exercises involve little plantarflexion (pointing the toes), they reason those lifts would be suboptimal for the calves. Nobody has tested it directly yet — but for those muscles, compounds alone probably aren't enough.
The one study among the seven in trained lifters, by de França and colleagues, pointed the same way. Twenty resistance-trained men followed a compound-only program, or the same program with curls and extensions added, for eight weeks. Neither measure of arm circumference differed significantly between the groups [1]. In that study, compounds were enough for total size — measured as arm circumference, not with a scan.
The evidence boundary draws itself: for the arm muscles these studies measured, total growth came out similar. For the hamstrings and calves, nobody has run the comparison. And for regional shape — the question Brandão's data opens — the evidence is early, sharp, and worth asking about every muscle in the program.
The question was not only 'do curls build more muscle?' It was also 'which part of the muscle do you want to grow?'
What Seven Studies Can and Can't Prove
The researchers who ran this meta-analysis policed their own inputs. They excluded studies from one research group over concerns about data integrity — a team voluntarily pulling ingredients they weren't confident in. Seven studies isn't a mountain of evidence. The researchers are the first to say so, describing their own estimate as "relatively imprecise."
But those seven studies sit close together: the authors found no appreciable heterogeneity between them. And that's not because only the agreeing ones got published — the funnel plot looked symmetrical, with no suggestion of publication bias.
The regional finding — the shape data from Brandão — rests on one study examining one muscle. Whether the same differential growth occurs in other muscles with other exercises remains understudied. The similar-size finding stands on seven studies and an estimate the authors call relatively imprecise. The shape divergence stands on one study and a great deal of promise.
The Choice That Was Always Yours
For total size, compounds are enough. For regional shape, isolation can target parts of a muscle that compounds grow less. Neither side was wrong. The minimalist got the efficiency they claimed. The maximalist got the shape they sensed. Now both know more about what each option gives them.
If exercise selection changes the shape of the triceps at the regional level, does it change the shape of the quad too? Six of seven studies covered the upper body. For any lifter whose legs are part of the program, the lower body version of that question still needs an answer.
The evidence redraws the line between efficiency and specificity. Compound-only programs cover total hypertrophy — across seven pooled studies, adding isolation made a trivial difference. Adding isolation isn't a growth requirement; it's a tool for targeting specific parts of a muscle that compounds may grow less.
What other research found
What this means for you
Your efficiency instinct is backed by seven pooled studies — compound exercises built similar total muscle to programs that added isolation work. What the data also shows: different exercises can grow different regions of the same muscle more. One study found that bench press and extensions grew opposite heads of the triceps, with dramatically different magnitudes. If specific muscle heads matter to your aesthetic goals, it's worth checking which regions your program emphasizes — and which it leaves behind. On total size, the pooled evidence sides with you, though the authors call the estimate relatively imprecise. The regional question is newer, and worth asking.
The extra time you spend on curls and extensions did not show up as clearly more total muscle across the seven pooled studies. What isolation can do: target parts of a muscle that compounds grow less. Studies where the added isolation work meant more total sets came out much like studies where volume was matched. In the one study that measured the heads of a muscle, the triceps extension grew the long head far more than the bench press did. That distinction changes the question: not 'should I add these?' but 'do I care where the growth lands?'
Most of the research tested beginners. That raises a fair question for experienced lifters: does the finding hold for people who've been training seriously? One study tested that directly. Twenty men with at least two years of continuous training did either a compound-only or compound-plus-isolation program for eight weeks. No significant difference in arm circumference between the groups. For total size, compounds appear sufficient even when the muscle has years of training history. In that one study, the total-size finding held. The shape question is open — nobody has tested whether exercise selection changes regional architecture in trained muscle the way it did in recreationally active men.
Before you change anything
Who this evidence covers: young men, mostly without training experience, performing upper body exercises with moderate loads (8 to 15 reps per set). Six of seven studies trained the upper body; the meta-analysis looked at the arms and legs only, not the chest.
Who it doesn't cover yet: women (no studies), adolescents (no studies), older adults (one study, but the protocol wasn't effective for either group), the hamstrings and calves (no comparisons exist; the one leg study measured whole-thigh volume), and heavy or light loading schemes.
The total-size finding is consistent but imprecise: seven studies, no appreciable heterogeneity, no signs of selective publishing, and results that did not meaningfully change when two influential comparisons from one study were removed. The researchers themselves call the estimate 'relatively imprecise' — and single studies did not all point the same way.
The regional-shape finding is early and compelling: one study, one muscle, but with dramatic differences between exercise types. Whether the same pattern holds for other muscles, other populations, and longer training periods hasn't been tested.
Rosa's data is largely an upper-body story. For any lifter whose legs are part of the program, that leaves a question: if exercise selection changes the shape of the triceps at the regional level, does it change the shape of the quad too? A different study in this cluster tested exactly that — comparing leg press to leg extension over eight weeks. The answer involves the front of the thigh and the outside of the thigh growing in opposite directions.
What This Study Found
All findings from this paper, in plain language.
- Isolation and compound exercises produced similar total muscle growth across seven studies and ten direct comparisons.
- The researchers found no appreciable heterogeneity between the seven studies, though not every study pointed the same way.
- Publication bias testing showed no evidence of selective reporting, with a symmetrical funnel plot.
- Studies where isolation added extra sets came out much like studies with matched volume — a trivial difference, though the range around it was relatively wide.
- Whether a study compared isolation with compounds directly or added isolation on top of compounds made no meaningful difference to the result.
- Within the triceps, different exercises grew different heads more — extensions grew the long head over eight times more than bench press, the bench press the lateral head.
- Six of seven studies tested upper body — lower body evidence is severely limited to a single thigh measurement.
- No study has compared compound and isolation exercises for hamstrings or calves, and indirect evidence suggests compounds alone fall short for these muscles.
- The evidence comes mostly from young untrained or recreationally trained men — only one study tested trained lifters, and none exclusively tested women or adolescents.
- All included studies scored moderate to good quality on methodological assessment, with a median score of 22 out of 29.
- Removing two influential comparisons from one study did not meaningfully change the overall result.
- The researchers conclude that compounds alone achieve similar hypertrophy with less training time, but combining exercise types may help target individual muscle regions.