Steak Salad with Blue Cheese

What this is

Sliced steak over leaves with blue cheese, walnuts and a sharp dressing. Fifteen minutes, and it is the meal that proves a salad can be dinner.

It exists because a steak on its own is not a meal and a salad on its own is not either, and the combination handles both problems.

The walnut evidence here is the most rigorous piece of research in this entire recipe collection, and it is worth reading for how honestly it reports.

Ingredients

For 2 servings

1 steak, sirloin or rump, about 300 g 120 g mixed leaves: rocket, watercress, little gem 60 g blue cheese, crumbled 40 g walnuts, roughly broken 1 pear or apple, thinly sliced Half a red onion, very thinly sliced

Dressing 3 tbsp olive oil 1 tbsp red wine vinegar 1 tsp Dijon mustard 1 tsp honey Salt and black pepper

Optional: cherry tomatoes, cooked beetroot, a soft-boiled egg

Method

1. Take the steak out of the fridge 30 minutes ahead and pat it completely dry. Season generously just before cooking.

2. Toast the walnuts in a dry pan for 3 minutes until fragrant. Watch them; they go from toasted to burnt in seconds. Set aside.

3. Get a heavy pan properly hot. Sear the steak 2 to 3 minutes a side for medium rare on a 2 cm cut.

4. Rest the steak 5 to 8 minutes. This is not optional. Cutting early loses the juice onto the board rather than into the meat.

5. Whisk the dressing until emulsified.

6. Toss the leaves, onion and pear with most of the dressing.

7. Slice the steak against the grain, lay it over the salad, and scatter the blue cheese and walnuts. Spoon the remaining dressing over the meat.

8. Serve immediately, while the steak is still warm and the leaves have not wilted.

The walnut meta-analysis

This is the most rigorous evidence in the collection and it deserves reporting in full, including the parts that found nothing.

A GRADE-assessed systematic review and dose-response meta-analysis searched major databases to August 2025 for randomised controlled trials of walnut intake, examining triglycerides, total cholesterol, LDL and HDL cholesterol, and apolipoproteins A1 and B.

The pooled analysis covered 49 randomised controlled trials, 57 effect sizes, and 4,611 participants.

Walnut consumption significantly reduced triglycerides, with a weighted mean difference of 6.24 mg/dL and a confidence interval from 9.49 to 2.99.

And here is what did not move. Apolipoprotein B showed no significant change, at p = 0.145. Apolipoprotein A1 showed a weighted mean difference of zero, with a confidence interval from negative 0.01 to 0.00 and p = 0.573.

Zero. Confidence interval essentially on zero. Reported plainly in the same abstract as the positive finding.

That is what good evidence reporting looks like, and it is why this study is worth trusting more than one that only tells you about the thing that worked.

The authors are also candid about why the review was needed: several clinical trials had examined walnuts and lipids, results have been inconsistent, and evidence on apolipoproteins specifically was limited.

So the honest summary: walnuts moved triglycerides by a modest, real amount across 49 trials, and did not move two apolipoprotein markers at all. Forty grams in a salad is not a treatment, and this is a good reason to eat walnuts rather than a claim about your bloodwork.

Blood lipids and cardiovascular risk are matters for a clinician. Nothing here is medical advice.

On the steak being the protein

The animal against plant protein argument has a controlled feeding trial behind it that is worth knowing about.

Sixty-two overweight adults with metabolic syndrome consumed a healthy American diet for two weeks, then were randomly allocated to one of three diets: a modified DASH diet rich in plant protein at 18 percent protein with two thirds from plant sources, a modified DASH diet rich in animal protein at 18.4 percent with two thirds from animal sources, or a moderate-protein diet at 27 percent protein with two thirds animal.

The diets were compared across three phases of energy balance, beginning with five weeks of controlled weight maintenance where all foods were provided.

All foods provided. That is the gold standard in dietary research and it is expensive and rare, because it removes the guesswork of asking people what they ate.

The reasoning behind the trial is a fair point: food-based dietary patterns emphasising plant protein are recommended for treating metabolic syndrome, but the contribution of plant protein to total protein in those diets is proportionally less than that of animal protein. In other words, the diets being recommended as plant-forward still lean on animal protein more than the label suggests.

This article reports the trial design rather than its outcome figures, which are the substance of the paper.

Sixty-two overweight adults with metabolic syndrome, a specific clinical population. Nothing here is medical advice, and metabolic syndrome is a matter for a doctor.

Why this combination of ingredients holds up

A narrative review looked at foods and cognition across the lifespan, and this salad happens to contain several of the categories it examined.

The review synthesised evidence on plant- and animal-derived foods and cognitive outcomes from early development through older adulthood, focusing on eight major categories of brain-supportive foods: dairy, eggs, seafood, lean meat, berries, leafy green vegetables, nuts and whole grains.

This salad contains four of those eight.

The stated gap is interesting: while plant-based brain foods have been emphasised, the contributions of animal-sourced foods to neurodevelopment and cognitive performance remain underexplored. So the review deliberately covers both sides rather than the fashionable one.

It reports that nutrient-dense foods, including both animal-sourced and plant-based, support cognitive outcomes across the life course, with eggs, meat and nuts linked to improved neurodevelopment in early childhood.

Credit where it is due: the review also considers nutrient bioavailability, dietary patterns, and the structural, socioeconomic and environmental factors influencing access to these foods. That last point matters. Telling people to eat well is easy; whether they can afford or reach these foods is a separate question the review takes seriously.

This is a narrative review synthesising evidence across life stages rather than a trial, and cognitive outcomes are influenced by far more than diet.

Storage and swaps

This does not keep. Assemble and eat it. Dressed leaves are finished within twenty minutes.

Cook the steak ahead if you must and serve it cold and thinly sliced, which is genuinely good and a legitimate way to use leftovers.

Swap blue cheese for feta, goat cheese or shaved parmesan if blue is not for you. It is a strong ingredient and the salad works without it.

Swap walnuts for pecans or hazelnuts. Toast them either way; untoasted nuts in a salad taste of nothing.

Swap steak for leftover roast chicken, halloumi or a tin of chickpeas crisped in a hot pan.

And rest the meat. Every time somebody complains their steak was dry, this is why.