Table 2.

Description of all selected articles.

References	Type of study	Population (n, age, sex, nutritional status, diagnosis)	Intervention (KDT type, duration)	Ketosis assessment	Outcomes	Dropout	Quality (MMAT*)
Di Lorenzo et al. (29), Italy	Prospective observational study	n = 96, 18–65 years old, all F, OW, migraine with or without aura. Controls: OW F without headache. Baseline migraine attacks per month (n ± SD): 2.91 ± 1.73.	6 months on a dietetic plan. —Participants (n = 45) on 4 weeks VLCKD based on food replacements, 4 weeks transitional diet with nutraceutical integrators, 4 weeks transitional diet without nutraceutical integrators, and the residual period standard diet. - Controls (n = 51): 6-month low-calorie standard diet.	Twice a day urine ketone dipstick tests: 4–10 mmol/L** during the VLCKD.	VLCKD: frequency and tablet use reduced after the first month of the diet (p < 0.0001). Transition period worsening, despite being improved compared with the baseline. Control: number of days with headaches and tablet intake reduced from month 3 (p < 0.0001), attack frequency reduced at month 6 (p < 0.0001).	N = 15 (16%) abandoned the study (9 at the third and 6 at the sixth month).	***
Di Lorenzo et al. (30), Italy	Open-label controlled interventional study	n = 18, 19–54 years old, 16F/2M, NW and OW, migraine with or without aura. Controls: healthy volunteers with comparable BMI. Baseline migraine attacks per month (n ± SD): 4.40 ± 2.70.	BMI ≥ 25: 4-week VLCKD. BMI < 25: 4-week MAD supplemented with lipid powder (MCT, LCT, and omega-3) with nutraceutical integrators.	Daily urine ketone dipstick test confirmed the presence of ketosis (not mentioned in the numeric results).	Reduction in attack frequency and duration (p < 0.001). No change to the first SSEP and VEP block of responses, but normalization of the interictally reduced VEP and SSEP (p < 0.01) habituation during the subsequent blocks.	n = 07 (28%) did not fulfill primary inclusion criteria.	**
Di Lorenzo et al. (33), Italy	Prospective open-label single-arm clinical trial	n = 18, 25–55 years old, 7F/11M, NW, cluster headache. Baseline migraine attacks per month (n ± SD): 108.17 ± 81.71.	12-week MAD supplemented with lipid powder (MCT, LCT, and omega-3) with nutraceutical integrators.	Not mentioned.	Full resolution of headache in 11/18 patients, at least 50% reduction in 4/18 patients.	n = 00.	***
Di Lorenzo et al. (34), Italy	Prospective open-label single-arm clinical trial	n = 18, >18 years old, 16F/2M, BMI 26.7 ± 4.6 kg/m2, episodic migraine without aura. Controls: HVs with a comparable BMI. Baseline migraine attacks per month (n ± SD): 4.70 ± 2.50.	4-week MAD supplemented with lipid powder (MCT, LCT, and omega-3) with nutraceutical integrators (ratio 1.7–2: 1).	Urine ketone dipstick on the day of testing >0.5.	Patients after KDT exhibited a reduction in attack frequency (p < 0.001), duration (p = 0.001), and disability (p < 0.001); normalization of the intercritical pain-related evoked potential habituation; no change in nBR; no change in BMI; HVs exhibited a physiologic habituation in the N-P amplitude slope of PREP, and both in homolateral and contralateral nBR.	n = 4 (18%) later determined that did not fulfill the primary inclusion criteria.	**
Di Lorenzo et al. (32), Italy	Single-center randomized double-blind controlled crossover phase 2 trial	n = 35, 18–65 years old, 29F/7M OW or obese, migraine with or without aura, prediabetes. Baseline migraine attacks per month (n ± SD): 4.83 ± 2.01.	4-week periods on a VLCKD or VLCnKD in a randomized order.	Urine ketone dipstick. 75.9% of patients on VLCKD positive (range 0.5–4.5 mmol/L).	4-week period VLCKD, despite inducing similar weight loss and glycemic profile, was significantly more effective than VLCnKD in preventing migraine attacks, as evidenced by a decrease in the frequency of migraine days and attacks and a >50% response rate.	n = 6 (17%) due to the excessive difficulty of the VLCnKD.	*****
Valente et al. (36), Italy	Retrospective observational study	n = 23, 47.22 ± 15.21 years old, 22F/1M, 10 NW, 8 OW, 5 obese, migraine. Baseline migraine attacks per month (n ± SD): 12.50 ± 9.50.	3 months on KDT tailored to the patient's characteristics: n = 4 VLCKD, n = 5 cKDT 2:1, n = 5 cKDT 1.5:1, n = 8 cKDT 1:1, n = 1 cKDT 0.5:1.	Some patients had blood measurement, but not reported because it was not systematically collected.	Reduction in monthly headache days (12.5 ± 9.5 vs. 6.7 ± 8.6; p < 0.001). Reduction in days of acute medication intake (11.06 ± 9.37 vs. 4.93 ± 7.99; p = 0.008).	n = 10 (30%) due to poor compliance, reported inefficacy, excessive weight loss, lost at follow-up without known reasons.	***
Bongiovanni et al. (31), Italy	Open-label single-arm clinical trial	n = 23, 18–57 years old, 21F/2M. Median BMI 26.5 (19.7–42.2), refractory chronic migraine. Baseline migraine attacks per month (days/range): 30/12–30.	3 months: BMI >30 kg/m2: VLCKD (800 kcal) BMI ≥ 25 kg/m2: LC (1,100–1,300 kcal) BMI < 25 kg/m2: LC (1,500–1,700 kcal) 1 month on transition phase: CHO increase: 30 g every week up to 150 g. 2 months on maintenance phase: LGIT.	Not measured.	Reduction in duration: hours per day (p < 0.0016) and days per month (p < 0.0001). Reduction in intensity of pain of each migraine episode (p < 0.0024). Reduction of doses of analgesics taken in a month. Decrease in weight and BMI (p < 0.0001).	n = 15 (39%) without mention of reason.	***
Haslam et al. (35), Australia	Pilot randomized controlled crossover trial	n = 16, >18 years old, 14F/2M mean BW 77.5 ± 15.9 kg mean body fat 35.6 ± 7.7% Migraine Baseline migraine attacks per month: not mentioned.	12-weeks divided into two 4-week dietary intervention periods interspersed with one washout period. cKDT: 3:1 ratio; lower ratios in order to aid compliance and retention; anti-headache diet: excluding dietary-related migraine triggers.	Urinary ketosis was measured for 81% of participants on 18 out of 28 days; average ketone level 7.2 mmol/L across the 28 days (range 2.0–14.0 mmol/L).	There were no statistically significant differences in migraine frequency, severity, or duration.	n = 10 (38%) due to being lost at follow-up or discontinued intervention (diet too difficult) or commenced another diet.	***
Putananickal et al. (37), Switzerland	Single-center randomized placebo-controlled double-blind crossover trial	n = 32, 18–65 years old, 37F/3M, BMI mean 23.96 ± 4.33 kg/m2. Episodic migraine. Baseline migraine attacks per month (n ± SD): 7.50 ± 2.60.	12-week treatment periods on exogenous DL-Ca-Mg-BHB supplement divided into 3 daily servings; placebo: mannitol. Ketone supplement and placebo were administered with glucose-free syrup to overshadow taste difference. No dietary intervention.	Blood ketone levels were assessed using a portable point-of-care blood ketone meter. —BHB after 40 min intake: 0.40 mmol/L (0.30, 0.50 mmol/L).	No beneficial effect in migraine frequency or intensity during BHB treatment.	n = 3 (22%) withdrawal of consent or due to intolerable side effects.	*****
Valente et al. (36), Italy	Retrospective observational study	n = 23, 47.22 ± 15.21 years old, 22F/1M, 10 NW, 8 OW, 5 obese, migraine. Baseline migraine attacks per month (n ± SD): 12.50 ± 9.50.	3 months on KDT tailored to the patient's characteristics: n = 4 VLCKD, n = 5 cKDT 2:1, n = 5 cKDT 1.5:1, n = 8 cKDT 1:1, n = 1 cKDT 0.5:1.	Some patients had blood measurement, but not reported because it was not systematically collected.	Reduction in monthly headache days (12.5 ± 9.5 vs. 6.7 ± 8.6; p < 0.001). Reduction in days of acute medication intake (11.06 ± 9.37 vs. 4.93 ± 7.99; p = 0.008).	n = 10 (30%) due to poor compliance, reported inefficacy, excessive weight loss, lost at follow-up without known reasons.	***
Lovati et al. (38), Italy (38)	Open-label single-arm clinical trial	First study: KDT group: n = 13, 36.9 ± 12.9 years old 11F/2M, BMI: 29.1 ± 5.4. LC group: n = 8, 50.9 ± 10.8 years, all F, BMI 28.9 ± 6.3. Baseline migraine attacks per month (n ± SD): 19.10 ± 6.50. Second study: KDT group: n = 26, 24F/2M. LC group: n = 6, 5F/1M.	3 weeks: KDT: normocaloric or hypocaloric MAD with MCT (10 g/d). LC: < 40% of CHO.	Carried out only in Study 1: Urinary ketone levels were measured in all patients, independently if KDT or LC: ketones were labeled as absent (n = 97), low (+; n = 54), moderate (++; n = 135), and high (+++; n = 70).	Based on the first study: relationship between ketone production and effect on headache; the higher the ketones, the lower the migraine frequency (chi-square p = 0.0073).	First study: n = 5 (19%) retired the consent after learning about the dietary changes. Second study: n = 0.	***

KDT, ketogenic diet therapies; BW, body weight; BMI, body mass index; OW, overweight; NW, normal weight; F, female; M, male; SD, standard deviation; CHO, carbohydrates; LC, low-carb diet; VLCKD, very low-calorie ketogenic diet; VLCnKD, very low-calorie non-ketogenic diet; MAD, modified Atkins diet; LGI, low glycemic index diet; cKDT, classical ketogenic diet therapy; BHB, beta-hydroxybutyrate; SSEPs, somatosensory-evoked potentials; VEP, visual-evoked potentials; MCT, medium-chain triglycerides; LCT, long-chain triglycerides; HVs, healthy volunteers; PREP, pain-related evoked potentials; nBR, nociceptive blink reflex; N-P, negative-positive.

^*MMAT: Mixed Methods Appraisal Tool system (27): from ^* (1 star = lower quality) up to ^***** (5 stars = higher quality).

^**Conversion for ketones measured in urine (acetoacetate): 10 mg/dL is equivalent to 1 mmol/L.