3.1. Patient data
We enrolled 92 pediatric patients (55 males; 37 females) aged 3 months to 10 years (29.26 ± 24.18 months) (Fig. 1). Patients were on the KD for 1–12 months (mean KD duration 6.68 months). The average duration of patients’ epilepsy and treatment with anticonvulsant medication ranged from 1 to 120 (interquartile range [IQR] 4.5, 12, 24) months prior to KD commencement.
Twenty-three patients were not included the study. Eight opted to undergo epilepsy surgery or other treatment without ketone production after a routine KD examination, 6 discontinued within 3 days due to food refusal, 5 discontinued due to allergy, and 4 withdrew from ketone production within 1 week at their parents' request. The 92 patients included in this study experienced 1–5 (2.37 ± 1.23) types of seizures before their treatment with KD therapy: 60.87% experienced spasms, 44.57% partial, 40.22% generalized tonic-clonic, 36.96% atonic, 35.87% clonic, 31.52% tonic, 26.09% absence seizures, and 8.70% myoclonic seizures. Seizure etiology was determined in 30.43% of patients. In this group, 23.91% of patients were found to have a known genetic etiology, while 6.52% of patients had a structural abnormality visible on MRI. Of all of the seizures, 69.57% were of unknown etiology. The SCN1A gene mutation was observed in nine patients, and CDKL5 in two. Other patients had gene mutations such as in GRIN2A, STXBP1, ALG1, PDHX, POLG, GABRD, STAMBP, SLC6A1, and chrM-10158. The structural etiology was sequelae from intracranial hemorrhage or encephalitis. All children began KD therapy with a 2:1 ratio of fat and non-fat foods. Some patients adjusted the ratio, as allowed in the study protocol, while others (83.70%) adhered to the 2:1 KD ratio throughout the study period. Blood ketone levels ranged from 2 to 5 mmol/L. Prior to starting KD therapy, all patients had been treated with 1–10 (4.28 ± 0.19) types of AEDs (Table 1): 75% received valproate; 70.65%, topiramate; 70.65%, levetiracetam; 66.30%, glucocorticoids; and 55.43%, nitrazepam. If KD therapy was effective, the AED doses were gradually tapered one by one.
Table 1. Baseline characteristics
Characteristics
|
Values
|
Age at KD commencement (months)
Disease duration (months)
Sex (M/F)
|
3-110 (29.26±24.18)
1-120, IQR (4.5, 12, 24)
92 (55/37)
|
Classification of seizure types
|
n (%)
|
Generalized onset
Focal onset
Unknown onset
|
42 (45.65)
41 (44.57)
9 (9.78)
|
Etiology
|
n (%)
|
Genetic
|
22 (23.91)
|
Structural
|
6 (6.52)
|
Unknown
|
64 (69.57)
|
KD ratio
|
n (%)
|
1/1
|
1 (1.09)
|
2/1
|
77 (83.70)
|
3/1
|
3 (3.26)
|
4/1
|
9(9.78)
|
4.5/1
|
2 (2.17)
|
AEDs
|
n (%)
|
Valproate
Topiramate
Levetiracetam
Glucocorticoid
Nitrazepam
Oxcarbazepine
Clonazepam
Vigabatrin
Lamotrigine
Phenobarbital
Carbamazepine
Zonisamide
Clobazam
Intravenous immunoglobulin
Rapamycin
|
69 (75.00)
65 (70.65)
65 (70.65)
61 (66.30)
51 (55.43)
23 (25.00)
20 (21.74)
12 (13.04)
12 (13.04)
11 (11.96)
5 (5.43)
4 (4.35)
4 (4.35)
2 (2.17)
1 (1.09)
|
A total of 92, 68, 55, and 29 patients received KD treatment for 1, 3, 6, and 12 months, respectively. None exhibited an increase in seizure frequency. At 1, 3, 6, and 12 months, 72/92 (78.26%), 51/68 (75%), 44/55 (80%), and 27/29 (93.10%) of the patients were responders, respectively. Five patients who were not responsive at 3 months were responsive by 6 months (Table 2).
Table 2. KD Efficiency
Efficacy (n [%])
|
1 month
|
3 months
|
6 months
|
12 months
|
n
|
92
|
68
|
55
|
29
|
Seizure-free
|
10 (10.87)
|
10 (14.71)
|
8 (14.55)
|
9 (31.03)
|
Improvement (≥50% reduction)
|
62 (67.39)
|
41 (60.29)
|
36 (65.46)
|
18 (62.07)
|
No improvement (<50% reduction)
|
20 (21.74)
|
17 (25.00)
|
11 (20.00)
|
2 (6.90)
|
A total of 39 patients (42.39%) discontinued the diet for various reasons, such as lack of efficacy (71.79%), caregiver non-compliance (10.26%), child's refusal to eat (7.69%), death (7.69%), or side effects (2.56%) (Table 3). There were three cases of death: one due to sudden invasive Streptococcus pneumoniae infection and empyema during the 10th month of the KD, another due to sudden arrhythmia in the 1st month of the KD, without any convulsions on the day of death, and the third due to dehydration from diarrhea and vomiting, because he could not receive intravenous rehydration for several days, as his home was far from our hospital; this occurred during the 6th month of the KD.
Table 3. Reasons for discontinuing KD therapy
Reasons
|
n (%)
|
Lack of efficacy
Caregiver's non-compliance
Child's refusal to eat
Death
Side effects
|
28/39 (71.79)
4/39 (10.26)
3/39 (7.69)
3/39 (7.69)
1/39 (2.56)
|
Of the 92 patients, 81 reduced the number of AEDs (12 patients discontinued all AEDs), while 11 underwent no change in AEDs. Among the 81 patients who reduced their AEDs, 11 later increased their AEDs because their convulsions recurred after 6–12 months. All 92 patients had been treated with 1–10 (4.28 ± 0.19) types of AEDs before KD therapy. At the final follow-up, the mean number of AEDs was 0–4 (2.45 ± 0.17). The overall mean reduction in AEDs was 1.83 ± 0.14 (p < 0.001). The EEG and MRI findings also underwent significant changes (p < 0.001, p = 0.04) (Table 4).
Table 4. Change in EEG and MRI findings
|
Initial
|
|
Reviewed
|
χ2
|
p
|
|
Normal
|
Abnormal
|
|
Normal
|
Abnormal
|
EEG
|
2
|
90
|
|
15
|
62
|
13.88
|
<0.001
|
MRI
|
29
|
63
|
|
34
|
36
|
4.86
|
0.04
|
EEG reports were available for all patients. Spike counts were used to evaluate EEGs. Before the KD, the EEG was normal in only 2 of 92 patients; 97.83% had abnormal EEGs, indicating that 90/92 (97.83%) had epileptiform discharges and 58/92 (63.04%) had slowing waves. Those included 37 patients with hypsarrhythmia, with generalized slowing in 21 of them. After KD therapy for 3–6 months, the EEGs of 77/92 (83.70%) of the patients were retested; 15/77 (19.48%) had become normal. Epileptiform discharges ceased in 23/77 (29.87%) after ketogenesis, decreased in 44/77 (57.14%), and did not improve in 10/77 (12.99%) patients. The background of EEGs became normal in 45/77 (58.44%) after ketogenesis, improved in 22/77 (28.57%), and exhibited no improvement in 10/77 (12.99%) after ketogenesis.
MRI reports were available for all patients. Before the KD, the findings were abnormal in 63/92 (68.48%) patients and normal in 29/92 (31.52%). After KD therapy for 3–6 months, 70/92 (76.09%) patients underwent further MRI; abnormal findings were observed in 36/70 (51.43%) and normal findings were observed in 34/70 (48.57%) patients; further, 12/70 (17.14%) exhibited a worsened condition. Among these 12 patients, 6 changed from normal to abnormal. Further, 10 of these 12 had invalid KDs, and 2 underwent baseline MRI 6 months before ketogenesis, during which the seizures were frequent, up to dozens of times per day, suggesting that seizures might aggravate MRI. Wide extra-cerebral space, encephalatrophy, and encephalomalacia were the main imaging findings. The patients who responded effectively to the KD exhibited no aggravation due to MRI. With the exception of one patient with glucose transfer factor 1 deficiency, MRI revealed improvement in encephalatrophy after treatment with KD; the other patients exhibited no significant improvement in MRI findings.
Age (p = 0.45), duration of disease (p = 0.82), type of seizure (p = 0.25), etiology (p = 0.30), type of AED (p = 0.85), and KD ratio (p = 0.41) were not associated with treatment efficacy (Table 5).
Table 5. Factors associated with KD efficacy Hypothesis
Factor
|
OR
|
95% CI
|
z
|
p
|
Age
|
1.02
|
0.98-1.06
|
0.76
|
0.45
|
Disease course
|
0.99
|
0.95-1.04
|
-0.23
|
0.82
|
Seizure type
|
1.38
|
0.80-2.38
|
1.15
|
0.25
|
Etiology
|
1.89
|
0.57-6.26
|
1.04
|
0.30
|
AED type
|
0.85
|
0.62-1.17
|
-0.99
|
0.32
|
KD ratio
|
0.70
|
0.30-1.64
|
-0.82
|
0.41
|
Most parents in our study (78.26%) reported that their children had demonstrated progress in the intellectual, language, or movement components, but the increases of intelligence assessment were not significant (p > 0.05). However, 11.96% reported that their child did not exhibit changes in any of these measures of intelligence, and 9.78% of parents reported that they could not evaluate their child’s progress. Some patients were unable to complete the intelligence assessments due to disturbances in consciousness or their parents’ refusal to perform the evaluation; 55.43% of all patients completed the intelligence assessment and 35.29% (18/51) of the parents reviewed the assessment.
Complications of the KD included transient hyperlipidemia in 55.43% of patients, hyperuricemia in 2.17%, nephrolithiasis in 2.17%, and hypoalbuminemia in 1.09%. Transient hyperlipidemia and hyperuricemia were reduced to normal without any treatment. Nephrolithiasis disappeared after potassium citrate usage was increased. Hypoalbuminemia was cured by reducing the KD ratio (increasing the protein content).