The Ketogenic Diet & Your Gains: Study Shows Body Comp. Benefits (Fat ↓), but Decreased Muscle Gains, However...

Unfortunately, the authors do not report what exactly (foods) the subjects ate.

I am pretty sure you will already have seen the results of Salvador Vargas' recent study in the Journal of the International Society of Sports Nutrition elsewhere, there are good reasons for me to still address what the authors claim was the first study to determine "if following a KD hypercaloric diet would promote greater gains in fat free mass and fat loss during a hypertrophic training period in resistance-trained men" (Vargas 2018) in an individual post - not just, but also because the scientists were unable to confirm their hypothesis, that a "KD with caloric surplus in combination with RT in trained men would have a positive impact" on both body fat and lean body mass (LBM).
From a methodological perspective, the randomized, parallel arm, controlled, prospective study looks solid - even though the exclusion of anyone who consumes dietary supplements and the neither physique-, nor performance-based inclusion criteria of having more than 2 years of continuous experience in overload training may not necessarily represent what the average gymrat would think of when he/she hears about "resistance-trained individuals" - with an average age of 30 ± 4.5 years, a body weight of 76.7 ± 5.7 kg at a height of 177 ± 3.4 cm (BMI = 23.4 ± 2.2 kg/m²) and a body fat percentage of 15 and 17% in the NKD and KD group, respectively (the difference was statistically non-significant), the subjects came much closer to this concept than the average "trained" study participant who is, often in spite of having trained for years, both weak and fat... accordingly, it's not totally surprising that the control group, who continued on their self-designed programs and kept their regular non-ketogenic eating habits, gained almost as much lean mass as the NKD group.

Figure 1: Overview of training protocol. WK: Workout (microcycle); UL: Upper-Limb; LL: Lower-Limb; R: Rest; 30X: 3 s of eccentric contraction and explosive movement concentric (Vargas 2018).

Something similar can be said about the subjects' workout routine. As you can see in Figure 1, the subjects trained four times a week in form of a repeated upper-/lower-body split using a selection of classic resistance training exercises that ranged from bench presses to triceps dips (UL-day) to squats and machine calf raises (LL-day).

In combination with the three rest days, the 32 workouts all subjects performed over the course of the 8-week study triggered the expected improvements in body composition (=an improvement of the lean-to-fat-mass ratio) in all subjects irrespective of whether they had been randomized to the control diet, the ketogenic diet (KD), or the non-ketogenic diet (NKD).
A closer look at the actual study results does yet also show significant inter-group differences for both, the individual gains in lean mass and reductions in fat mass in the NKD and KD study groups:

Figure 2: Changes in parameters of body composition over the 8-week study on KD, NKD or control diet (Vargas 2018).

If we do the math, the diverging increases in lean and decreases in fat-mass amount to a higher (albeit not significantly) increase in body composition if the latter is defined by the subjects' body fat %, I calculated based on the absolute changes in body weight and fat mass.

There are things you should be aware of before jumping to conclusions

The first thing doesn't even require looking beyond the results of the study at hand: (1) the overall effect of the workout + diet combinations are marginal or small, (2) only the increase in lean mass in the NKD and the decrease of fat mass in the KD group were statistically significant with p < 0.05, and (c) the study which must be considered a small scale pilot clearly suffers from the lack of dietary control. The latter is particularly problematic because previous studies have observed significant appetite reducing effects with ketogenic diets. In the absence of rigorous dietary control, both, the increased fat loss and the decreased muscle mass could well be a result of the subjects' non-compliance with the standardized energy target of 39kcal/kg/d.

Figure 3: According to the scientists' calculation of effect sizes even the statistically significant effects of the KD on body fat and the NKD on lean mass were only "small" - an observation that's not uncommon in trained individuals.

What I think is yet even more important, though, is that using DXA, the alleged "gold standard" for body composition standards, in studies that involve one or multiple transitions from regular high-carb to ketogenic diets and vice versa are prone to yield unreliable results. I've discussed that at length in my dissertation to Wilson's often-criticized keto-gains study, a study that used a virtually identical KD as the study and hand and still yielded fundamentally different results... because Wilson et al. decided to (carb-)refeed their subjects before the final measurement. Whether that was a deliberate mistake, or not has not just been debated more than often enough, it's also irrelevant when it comes to the significance of the results of the study at hand.

What is relevant, however, is that comparing DXA data from high- vs. low-carb phases is akin to comparing apples and oranges; and this is particularly true if the standard 2-compartment model is used (cf. Toombs 2012 | note: the study at hand didn't report which method Vargas et al. used, but if they'd used the more accurate but still not infallible 4-component- aka 4C-model, I am sure the scientists would have highlighted that).
The small decrease in lean mass and, to a certain degree, the increased loss of body fat may thus be nothing but artifacts the measurement method generates when it's used on subjects on a diet that has repeatedly been shown to trigger both significant water- and glycogen-reductions (Nana 2012; Burke 2017) and thus physiological changes of which the researchers have highlighted previously that they could explain the "substantial changes in [DXA-based] estimates of body composition over the course of the day" Bone & Burke observed in previous studies (Bone 2016).

Figure 4: Relative changes in leg lean and fat mass vs. baseline following glycogen depletion and glycogen loading with and without creatine (Bone. 2017) - this particular figure was initially published in a SuppVersity article from 2016.

As highlighted in a 2016 SuppVersity Article, Bone and Burke observed highly significant increases in lean and fat mass within only 3 days when they fed well-trained cyclists a high(er = 2x more than baseline) carbohydrate diet (see data in Figure 2 | Bone 2017); the exact opposite, i.e. statistically significant and, if you falsely take them at face value, practically relevant reductions in both lean and fat mass, was observed by Bone & Burke as a result of the initial glycogen depletion in the same study.

A fair comparison using DXA data would require normalizing the groups' glycogen levels

In that, it is worth noting that Bone's results deviate from a previous study by Rouillier et al. who observed a smaller increase in body fat (compared to lean mass) and thus an overall decrease in body fatness in response to 3 days on a high carbohydrate diet in their 2015 study. Whether and to which extent the differential effects on the subjects' measured (not real) fat mass can be explained by methodological differences such as the lack of intense training and/or differences in the subjects' athleticism between the studies cannot be said for sure at the moment.

Figure 5: Let's be clear here, I don't advocate super-compensation strategies as they were used in Wilson's 2017 study week 11, as this would skew the data in the opposite way(s), i.e. exaggerate the lean mass gains on a ketogenic diet.

What we can say with reasonable confidence, though, is that you shouldn't put too much confidence into the accuracy/objectivity of the DXA data from the study at hand. Gold standard or not. DXA is so prone to be skewed by changes in glycogen and body water that t would require a glycogen normalization intervention before the final DXA scan - one that does not provide an unfair advantage to the keto group, though (see the previously discussed study by Wilson et al.)...

What? Oh, no, this does not mean that all previous studies on the effects of ketogenic diets on body composition are bunk. In view of the fact that >90% of them are weight loss trials, the vast majority of this study will have a certain degree of muscle glycogen depletion in both the keto and the control diet simply due to the lack of total energy. This is in contrast to the study at hand, where the non-ketogenic and the control diet were not just significantly higher in carbohydrates, but, in view of the lack of dietary control and the repeatedly observed appetite suppressant effects of ketogenic diets, probably also significantly higher in total energy intake. Against that background, it is reasonable to assume that any glycogen-/water-related differences in DXA-based estimates of body composition would be exasperated compared to studies that used high vs. low carb diets for weight loss purposes.
References:

• Bellissimo, Nick, and Tina Akhavan. "Effect of macronutrient composition on short-term food intake and weight loss." Advances in Nutrition 6.3 (2015): 302S-308S.
• Bone, Julia L., et al. "Manipulation of muscle creatine and glycogen changes dual x-ray absorptiometry estimates of body composition." Medicine & Science in Sports & Exercise 49.5 (2017): 1029-1035.
• Bone, Julia, and Louise M. Burke. "DXA estimates of body composition and carbohydrate loading." Annals of Nutrition and Metabolism 68.3 (2016b): 228-229.
• Burke, L. "Dual X-Ray Absorptiometry (DXA) for measurement of body composition in athletes: Experiences that underpin the importance of optimising the reliability of measurement." Journal of Science and Medicine in Sport 20 (2017): 76.
• Gibson, Alice A., et al. "Do ketogenic diets really suppress appetite? A systematic review and meta‐analysis." Obesity Reviews 16.1 (2015): 64-76.
• Nana, Alisa, et al. "Effects of daily activities on dual-energy X-ray absorptiometry measurements of body composition in active people." Medicine & Science in Sports & Exercise 44.1 (2012): 180-189.
• Nymo, Siren, et al. "Timeline of changes in appetite during weight loss with a ketogenic diet." International Journal of Obesity 41.8 (2017): 1224.
• Rouillier, Marc-Antoine, et al. "Effect of an acute high carbohydrate diet on body composition using DXA in young men." Annals of Nutrition and Metabolism 66.4 (2015): 233-236.
• Toombs, Rebecca J., et al. "The impact of recent technological advances on the trueness and precision of DXA to assess body composition." Obesity 20.1 (2012): 30-39
• Vargas, Salvador, et al. "Efficacy of ketogenic diet on body composition during resistance training in trained men: a randomized controlled trial." Journal of the International Society of Sports Nutrition 15.1 (2018): 31.