Efficacy, safety and tolerability of very low-calorie ketogenic diet in obese women with fibromyalgia: a pilot interventional study

Abstract

Introduction: Obesity can worsen fibromyalgia (FM) and very low-calorie ketogenic diet (VLCKD) is a potential therapeutic option for diseases that share clinical and pathophysiological features with FM. In this pilot interventional study, we investigated the effects of VLCKD in obese women with FM.

Methods: Female patients with FM and a body mass index (BMI) ≥ 30 kg/m² were eligible for VLCKD. The ketogenic phase (T0 to T8) was followed by progressive reintroduction of carbohydrates (T8 to T20). Changes in BMI, Fibromyalgia Impact Questionnaire (FIQ), Hospital Anxiety and Depression Scale (HADS), EuroQol 5D (EQ-5D) and 36-item Short Form Health Survey (SF-36) were evaluated. A change of 14% in FIQ was considered clinically relevant. The longitudinal association between BMI and patient-reported outcomes (PROs) was assessed using generalized estimating equations.

Results: Twenty women were enrolled. Two discontinued the intervention. The mean age of the 18 patients who reached T20 was 51.3 years and mean BMI was 37.2 kg/m². All patients lost weight during the first period of VLCKD and this achievement was maintained at T20. Mean BMI decreased from 37.2 kg/m² at T0 to 34.8 kg/m² at T4, 33.5 kg/m² at T8 and 32.1 kg/m² at T20 (p < 0.001). A significant reduction of mean FIQ from 61.7 at T0 to 37.0 at T4 and to 38.7 at T8 (p < 0.001) was observed and it was maintained at T20 with a mean FIQ of 39.1 (p = 0.002). Similar results were obtained for HADS, EQ-5D and SF-36. Analysing each participant, the reduction of FIQ was clinically meaningful in 16 patients (89%) at T4, in 13 (72%) at T8 and in 14 (78%) at T20. No significant association was observed between change in BMI and improvement of the PROs over time. Adverse effects were mild and transient. No major safety concerns emerged.

Conclusion: These are the first data on the efficacy of VLCKD in FM. All patients achieved improvement in different domains of the disease, which was maintained also after carbohydrate reintroduction. Our results suggest that ketosis might exert beneficial effects in FM beyond the rapid weight loss.

Clinical trial registration: This trial is registered on ClinicalTrials.gov, number NCT05848544.

Keywords: fibromyalgia; ketogenic diet; ketone bodies; obesity; pain.

Figure 1

Flowchart of the study protocol. A 4 weeks run-in period of free diet preceded the weight-loss program (Phase 0 – T-4 to T0). The ketogenic period was divided into three phases. During the first 4 weeks (Phase 1 – T0 to T4), patients were allowed to eat 4 to 6 protein preparations every day with an average daily caloric intake of 801 kcal. The next 4 weeks were divided into two 2 weeks periods (Phase 2 – T4 to T6 and Phase 3 – T6 to T8), in which one (Phase 2) or two (Phase 3) of the meal preparations were replaced by natural proteins, for an average daily caloric intake of 843 kcal. Carbohydrates were then progressively reintroduced, starting from foods with low glycemic index during the next 4 weeks (Phase 4 – T8 to T12) and then continuing with moderate (Phase 5 – T12 to T16) and high (Phase 6 – T16 to T20) glycemic index products. The average daily caloric intake gradually increased from 1,138 kcal in the fourth phase to 1,186 kcal in the fifth phase and 1,490 kcal in the sixth phase. The presence of ketosis was assessed at weekly intervals from T0 to T8 and then at T12 using urine strips.

Figure 2

Average daily caloric intake for each phase of the study. In the first phase, the average daily caloric intake was 801 kcal, with a macronutrient ratio of 11% carbohydrates, 39% fat, 44% proteins and 6% fibers. In the second and third phases, the average daily caloric intake was 843 kcal, with a macronutrient ratio of 10% carbohydrates, 40% fat, 44% proteins and 6% fibers. In the fourth phase, the average daily caloric intake was 1,138 kcal, with a macronutrient ratio of 25% carbohydrates, 34% fat, 32% proteins and 9% fibers. In the fifth phase, the average daily caloric intake was 1,186 kcal, with a macronutrient ratio of 26% carbohydrates, 33% fat, 32% proteins and 9% fibers. In the sixth phase, the average daily caloric intake was 1,490 kcal, with a macronutrient ratio of 48% carbohydrates, 28% fat, 20% proteins and 4% fibers.

Figure 3

CONSORT flow diagram. Of the 28 patients assessed for eligibility, 8 were excluded and 20 were enrolled in the study. Of these, 2 discontinued the intervention and 18 completed the study period and were included in the analysis.

Figure 4

Mean change in body mass index [BMI – (A)], Fibromyalgia Impact Questionnaire [FIQ – (B)], Hospital Anxiety and Depression Scale [HADS – (C,D)], EuroQol 5D [EQ-5D – (E,F)] and 36-item Short Form Health Survey [SF-36 – (G,H)] from T-4 to T20. Level of significance of each measurement compared with T0 is indicated on the plots: ns = non-significant; * = <0.05; ** = <0.01; *** <0.001. (A) Mean BMI was 37.2 ± 3.8 at T-4 and 37.2 ± 4.0 at T0 (p = 0.995). During the ketogenic phase, mean BMI significantly decreased to 34.8 ± 3.9 at T4 (p < 0.001) and to 33.5 ± 3.6 at T8 (p < 0.001). Compared with FIGURE 4 (Continued)baseline values, the decrease in mean BMI was also significant during the carbohydrate reintroduction period, with 32.9 ± 3.5 at T12 (p < 0.001), 32.3 ± 3.4 at T16 (p < 0.001) and 32.1 ± 3.4 at T20 (p < 0.001). (B) Mean FIQ was 67.6 ± 13.9 at T-4 and 61.7 ± 22.2 at T0 (p = 0.124). Compared to T0, mean FIQ improved during the ketogenic period to 37.0 ± 23.0 at T4 (p < 0.001) and to 38.7 ± 21.6 at T8 (p < 0.001), with significant changes to 44.4 ± 25.0 observed also at T12 (p = 0.012), to 35.8 ± 23.7 at T16 (p < 0.001) and to 39.1 ± 25.9 at T20 (p = 0.002). (C) Mean HADS-A was 11.3 ± 4.1 at T-4 and 11.1 ± 4.0 at T0 (p = 0.633). Compared to T0, HADS-A significantly improved to 7.1 ± 3.7 at T4 (p < 0.001) and to 7.8 ± 4.1 at T8 (p < 0.001). Furthermore, HADS-A decreased to 8.3 ± 4.0 at T12 (p = 0.001), to 7.4 ± 3.4 at T16 (p < 0.001) and to 8.2 ± 3.5 at T20 (p = 0.004). (D) Mean HADS-D was 10.0 ± 3.5 at T-4 and 9.6 ± 3.7 at T0 (p = 0.571). Compared to T0, mean HADS-D significantly improved to 8.1 ± 3.8 at T4 (p = 0.021), to 7.9 ± 4.5 at T8 (p = 0.030) and to 6.9 ± 4.5 at T16 (p = 0.005). At T12 and T20, a non-significant reduction, respectively, to 8.6 ± 4.6 (p = 0.219) and to 8.0 ± 4.6 (p = 0.116) was observed. (E) Mean EQ-5D utility score was 0.61 ± 0.23 at T-4 and 0.64 ± 0.21 at T0 (p = 0.469). Compared to T0, mean EQ-5D utility score significantly increased to 0.78 ± 0.12 at T4 (p = 0.010), to 0.79 ± 0.12 at T8 (p = 0.003) and to 0.81 ± 0.13 at T16 (p = 0.002), while there was no significant difference between the baseline score and the values of 0.70 ± 0.24 (p = 0.333) and 0.73 ± 0.31 (p = 0.295) observed, respectively, at T12 and at T20. (F) Mean EQ-5D VAS score was 45.4 ± 19.0 at T-4 and 42.0 ± 23.7 at T0 (p = 0.297). Compared to T0, mean EQ-5D VAS score significantly improved to 59.1 ± 23.0 at T4 (p = 0.002), to 60.6 ± 22.5 at T8 (p = 0.009), to 57.0 ± 24.8 at T12 (p = 0.038), to 65.9 ± 25.2 at T16 (p < 0.001) and to 63.5 ± 30.1 at T20 (p = 0.022). (G) Mean SF-36 MCS was 34.3 ± 16.7 at T-4 and 33.5 ± 16.6 at T0 (p = 0.849). Compared to T0, mean SF-36 MCS significantly increased to 52.4 ± 24.0 at T4 (p < 0.001), to 50.1 ± 21.5 at T8 (p < 0.001), to 47.0 ± 26.6 at T12 (p = 0.010), to 55.5 ± 25.9 at T16 (p < 0.001) and to 49.8 ± 26.3 at T20 (p = 0.002). (H) Mean SF-36 PCS was 30.3 ± 14.8 at T-4 and 30.5 ± 17.0 at T0 (p = 0.942). Compared to T0, mean SF-36 PCS significantly increased to 47.3 ± 19.5 at T4 (p < 0.001), to 47.7 ± 20.0 at T8 (p < 0.001), to 41.6 ± 22.2 at T12 (p = 0.015), to 50.5 ± 20.9 at T16 (p < 0.001) and to 47.4 ± 23.8 at T20 (p = 0.009).

Figure 5

Change in FIQ in individual patients from the baseline visit to T4 (A), T8 (B) and T20 (C). (A) At T4, FIQ improvement was observed in 17 patients (94%) and it was clinically meaningful in 16 (89%); FIQ worsening was observed in 1 patient (6%) and it was not clinically meaningful. (B) At T8, FIQ improvement was observed in 15 patients (83%) and it was clinically meaningful in 13 (72%); FIQ worsening was observed in 3 patients (17%) and it was clinically meaningful in 2 (11%). (C) At T20, FIQ improvement was observed in 16 patients (89%) and it was clinically meaningful in 14 (78%); FIQ worsening was observed in 2 patients (11%) and it was clinically meaningful in 1 (6%).