Keywords low-grade inflammation - ketamine - antidepressant effect - antisuicidal effect
Introduction
Low-grade inflammation (LGI) is a systemic condition characterized by the chronic
subclinical production of inflammatory factors, including C-reactive protein (CRP).
LGI has been implicated in the pathogenesis of various noncommunicable diseases,
such as metabolic diseases, cardiovascular diseases, degenerative disorders, and
severe mental disorders [1 ]
[2 ]
[3 ]. In a meta-analysis involving 13,541 patients with depression and
155,728 control individuals, 27% of the patient cohort had LGI (i. e., CRP
level≥3 mg/L) [1 ]. In patients with major
depressive disorder, LGI has been associated with adverse clinical profiles, such
as
an elevated rate of relapse, worse cognitive deficits, and a poor response to
antidepressants [4 ]
[5 ]
[6 ]. We previously reported an association between LGI and an elevated
risk of treatment-resistant depression (TRD), defined as the lack of response to at
least two distinct antidepressants administered at adequate dosages for appropriate
durations [4 ].
Low-dose ketamine exerts rapid antidepressant and antisuicidal effects in patients
with TRD and major depressive disorder with suicidal ideation [7 ]
[8 ]
[9 ]. Low-dose ketamine also
increases the level of brain-derived neurotrophic factor and promotes synaptogenesis
by modulating the mammalian target of the rapamycin pathway through the blockade of
the N-methyl-D-aspartate receptor; evidence also suggests that the anti-inflammatory
effect of low-dose ketamine mediates its antidepressant and antisuicidal effects
[10 ]
[11 ]
[12 ]. We previously demonstrated a reduction in the tumor necrosis
factor-α level 40 min after a single infusion of 0.5 mg/kg ketamine; this reduction
was associated with reduced total scores on the Montgomery-Asberg Depression Rating
Scale (MADRS) [12 ]. In a preclinical
murine model of postoperative depression, esketamine attenuated the nuclear factor
κB pathway – a prototypical proinflammatory signaling pathway – in the prefrontal
cortex [13 ]. Yang et al. reported that
proinflammatory cytokines, such as interleukin (IL)-6 and CRP, may serve as
biomarkers of treatment response to low-dose ketamine in patients with TRD [14 ]. However, whether the antidepressant
and antisuicidal effects of low-dose ketamine among patients with TRD differ between
those with and without LGI remains unclear.
In this study, we reanalyzed data from three of our clinical trials, wherein 167
patients with TRD received low-dose ketamine infusions. Our goal was to clarify how
LGI influences the antidepressant and antisuicidal effects of ketamine in patients
with TRD. Considering the anti-inflammatory effect of ketamine, we hypothesized that
low-dose ketamine exerts both antidepressant and antisuicidal effects in patients
with TRD, regardless of the presence of LGI.
Methods
Participants and study protocols
In the present study, we reanalyzed the data from our three low-dose ketamine
infusion clinical trials [8 ]
[9 ]
[15 ]. Details of the clinical trial protocols have been published in
our previous papers and were illustrated in [Fig. 1 ]
[8 ]
[9 ]
[15 ]. Briefly speaking, in clinical trials 1 and 3, 71 patients with
TRD and 48 patients with TRD and strong suicidal ideation were randomized to a
single low-dose ketamine (0.5 mg/kg or 0.2 mg/kg in clinical trial 1 and
0.5 mg/kg in clinical trial 3) or placebo (normal saline in clinical trial 1 and
0.045 mg/kg midazolam in clinical trial 3) group, respectively. All participants
in clinical trials 1 and 3 were followed up to 28 days postinfusion [8 ]
[9 ]. In clinical trial 2 phase 1, 49 patients with TRD (n=33) or
treatment-resistant bipolar depression (n=16) received two infusions of
0.5 mg/kg ketamine on day 1 and day 4, respectively. Those who responded to
0.5 mg/kg ketamine in clinical trial 2 phase 1 were subsequently randomized to a
7-week treatment of D-cycloserine or placebo in phase 2 [15 ]. TRD was defined as the failure of
treatment response for at least two different antidepressants with adequate
dosage and treatment duration [8 ]
[9 ]. Treatment-resistant bipolar
depression was defined as a lack of response to at least two trials with
antidepressants or mood stabilizers with documented efficacy in bipolar
depression (lithium, lamotrigine, quetiapine, or olanzapine) in adequate doses
and treatment duration [15 ]
[16 ]. Depressive symptoms were rated
using the 17-item Hamilton Depression Rating Scale (HDRS) and MADRS in three
clinical trials [8 ]
[9 ]
[15 ]. The Maudsley staging method for the level of treatment
refractoriness was also assessed [17 ].
In the current study, exclusion criteria included major medical (i. e.,
diabetes, autoimmune diseases, cardiovascular diseases, current infectious
diseases), neurological (i. e., stroke, epilepsy) diseases or a history of
alcohol or substance use disorders. Furthermore, those who used antibiotics or
immune modulators, such as non-steroidal anti-inflammatory drugs and steroids,
were excluded from our clinical trials. Regarding our research aim to examine
the role of LGI on the antidepressant effect of a single low-dose ketamine
infusion, we only analyzed the clinical data from day 1 to day 3 of three
clinical trials ([Fig. 1 ]). The
treatment response was defined as a≥50% reduction in overall depressive symptoms
at Day 2 postinfusion. Three clinical trials were registered in the UMIN
Clinical Trials Registry (UMIN000016985, R000027142-UMIN000023581, and
UMIN000033916). This study was performed in accordance with the Declaration of
Helsinki and was approved by the Taipei Veterans General Hospital Institutional
Review Board. Witten informed consents were provided by all participants.
Fig. 1 Study illustration. TRD: treatment-resistant major
depression; TRBD: treatment-resistant bipolar depression.
Measurement of C-reactive protein levels and definition of low-grade
inflammation
Fasting serum samples were collected between 8:00 AM and 10:00 AM in serum
separator tubes, clotted for 30 min, and stored at −80°C until use.
Enzyme-linked immunosorbent assay (ELISA) kits (R&D systems, Minneapolis,
MN, USA) were used to measure CRP levels for all participants. The vendor’s
instructions were followed for all assays. An ELISA plate reader with Bio-Tek
Power Wave Xs and Bio-Tek KC junior software (Winooski, VT, USA) was used to
measure the final absorbance of each sample of the mixture and analyze the same
at 450 nm. We considered the range specified in the vendor’s instructions as
standard. A linear regression R2 value of at least 0.95 was
considered a reliable standard curve. LGI was defined based on CRP levels≥3 mg/L
in the present study, which was compliant with the proposed cut-off point for
CRP levels on LGI in previous studies [1 ].
Statistical analysis
Continuous and nominal variables were analyzed through one-way analysis of
variance and Fisher’s chi-square tests, respectively, to assess the differences
in the demographic and clinical data between groups. After adjusting for age,
sex, body mass index (BMI), diagnosis, and psychiatric comorbidities, CRP level
(<3 or≥3 mg/L)-stratified generalized estimating equation (GEE) models with
the autoregressive method for correlations of repeated measures for the same
individual over time was used to examine the effect of ketamine on overall
depressive symptoms (total HDRS and MADRS scores) and specific suicidal symptoms
(HDRS item 3 and MADRS item 10 scores) during the study period (baseline to day
3) with the group (ketamine or placebo) as a between-patient factor, time as a
within-patient factor, and baseline depressive and suicidal symptoms as
between-patient predictors as well as all possible interactions, respectively.
Two-tailed p <0.05 was considered statistically significant. All data
processing and statistical analyses were performed using SPSS, version 17 (SPSS
Inc.).
Data availability
The datasets generated during and/or analyzed during the current study are not
publicly available due to Taiwan’s clinical trial ethical regulation but are
available from the corresponding author upon reasonable request.
Results
In all, 167 patients with TRD were administered a single infusion of low-dose
ketamine or placebo, including 46 in the LGI group and 121 in the non-LGI group
([Table 1 ]). Patients in the LGI
group were younger (p =0.004) and more obese (p <0.001) compared with
those in the non-LGI group ([Table 1 ]).
In addition, patients with TRD and LGI had higher rates of comorbid posttraumatic
stress disorder (p =0.010) and generalized anxiety disorder (p =0.037)
than those with TRD without LGI ([Table
1 ]). Other demographic and clinical characteristics, such as diagnosis,
duration of illness, levels of treatment refractoriness, and baseline overall
depressive (total HDRS and MADRS scores) and suicidal (HDRS item 3 and MADRS item
10
scores) symptoms, did not differ between groups (all p >0.05) ([Table 1 ]). In the non-LGI group, patients
in the low-dose ketamine group had a higher treatment response (HDRS: 41.9% vs.
8.6%, p <0.001; MADRS: 39.5% vs. 5.7%, p <0.001) than did those in
the placebo group ([Table 2 ]). However,
the treatment response rates did not differ between treatment groups (all
p >0.05) among patients with LGI ([Table 2 ]).
Table 1 Demographic and clinical characteristics between
groups.
Non-LGI group (n=121)
LGI group (n=46)
All sample (n=167)
p-value
Age (years, SD)
44.89 (12.92)
38.63 (10.78)
43.17 (12.65)
0.004
Sex (n, %)
0.438
Male
31 (25.6)
15 (32.6)
46 (27.5)
Female
90 (74.4)
31 (67.4)
121 (72.5)
BMI (SD)
22.31 (4.11)
27.52 (4.09)
23.75 (4.71)
<0.001
Diagnosis (n, %)
0.863
Major depressive disorder
105 (86.8)
41 (89.1)
146 (87.4)
Bipolar depression
16 (13.2)
5 (10.9)
21 (12.6)
Duration of illness (years, SD)
12.03 (8.38)
10.28 (7.98)
11.55 (8.29)
0.225
Infusion group (n, %)
0.454
0.5 mg/kg ketamine
70 (57.9)
27 (58.7)
97 (58.1)
0.2 mg/kg ketamine
16 (13.2)
6 (13.0)
22 (13.2)
0.045 mg/kg midazolam
15 (12.4)
9 (19.6)
15 (14.4)
Normal saline
20 (16.5)
4 (8.7)
20 (14.4)
MSM scores (SD)
8.88 (1.75)
9.37 (1.83)
9.02 (1.78)
0.116
Baseline CRP levels (mg/L)
0.70 (0.69)
6.60 (3.49)
2.32 (3.26)
<0.001
Clinical symptoms at baseline (SD)
Total HDRS scores
22.71 (4.38)
22.26 (3.53)
22.59 (4.16)
0.534
Total MADRS scores
35.51 (5.66)
35.52 (5.15)
35.51 (5.51)
0.992
HDRS item 3 scores
1.86 (0.81)
2.09 (0.81)
1.92 (0.81)
0.107
MADRS item 10 scores
3.13 (1.23)
3.50 (1.05)
3.23 (1.19)
0.075
Psychiatric comorbidities (n, %)
Panic disorder
49 (40.5)
23 (50.0)
72 (43.1)
0.297
PTSD
11 (9.1)
12 (26.1)
23 (13.8)
0.010
GAD
65 (53.7)
33 (71.7)
98 (58.7)
0.037
History of attempted suicide (n, %)
69 (57.0)
27 (58.7)
96 (57.5)
0.863
LGI: low-grade inflammation; SD: standard deviation; BMI: body mass index;
MSM: Maudsley Staging Method; CRP: C-reactive protein; HDRS: 17-item
Hamilton Rating Scale for Depression; MADRS: Montgomery-Åsberg Depression
Rating Scale; PTSD: post-traumatic stress disorder; GAD: generalized anxiety
disorder.
Table 2 Response rates between groups, as stratified by LGI
groups.
Non-LGI group (n=121)
LGI group (n=46)
Ketamine (n=86)
Placebo (n=35)
Ketamine (n=33)
Placebo (n=13)
Response rate at day 2 (n, %)
Based on HRDS
36 (41.9)
3 (8.6)
15 (45.5)
4 (30.8)
p-value
<0.001
0.510
Based on MADRS
34 (39.5)
2 (5.7)
14 (42.4)
2 (15.4)
p-value
<0.001
0.101
LGI: low-grade inflammation; HDRS: 17-item Hamilton Rating Scale for
Depression; MADRS: Montgomery-Åsberg Depression Rating Scale.
CRP level (<3 or≥3 mg/L)-stratified GEE models showed an antidepressant effect of
low-dose ketamine infusion only among patients with TRD without LGI (total HDRS
scores: group effect: p <0.001; total MADRS scores: group effect:
p <0.001) but not among those with TRD and LGI (p =0.535;
p =0.587) ([Fig. 2 ]). An
antisuicidal effect of low-dose ketamine was noted among patients with TRD without
LGI, measuring HDRS item 3 (group effect: p <0.001) and MADRS item 10
(p <0.001) ([Fig. 3 ]).
However, an antisuicidal effect of low-dose ketamine was only noted among patients
with TRD and LGI based on the measure of MADRS item 10 (group effect:
p =0.046) but not on that of HDRS item 3 (p =0.106) ([Fig. 3 ]).
Fig. 2 Trajectories of total HDRS and MADRS scores between groups,
stratified by CRP levels. CRP: C-reactive protein; HDRS: 17-item Hamilton
Rating Scale for Depression; MADRS: Montgomery-Åsberg Depression Rating
Scale.
Fig. 3 Trajectories of HDRS item 3 and MADRS item 10 scores between
groups, stratified by CRP levels. CRP: C-reactive protein; HDRS: 17-item
Hamilton Rating Scale for Depression; MADRS: Montgomery-Åsberg Depression
Rating Scale.
Discussion
Our findings revealed that among patients with TRD, only those without LGI responded
to a single infusion of low-dose ketamine. The antidepressant and antisuicidal
effects of low-dose ketamine were not observed in patients with LGI, potentially due
to the much stronger placebo effect in this group.
Immunometabolic depression is a subtype of depression characterized by the
convergence of LGI, metabolic dysregulation, and atypical depressive symptoms [18 ]. Findings from both our study and the
meta-analysis by Osimo et al. demonstrated that the prevalence of LGI (CRP≥3 mg/L)
in depression was approximately 27% [1 ].
However, our findings do not corroborate those of Park et al., who observed that
ketamine effectively mitigated both typical/melancholic and atypical symptoms of
depression [19 ]. However, 2 days after the
infusion of ketamine, its antidepressant effect was more pronounced against
typical/melancholic symptoms than against atypical symptoms [19 ]. The researchers further reported that
the immunometabolic and inflammatory mechanisms underlying atypical symptoms of
depression may be resistant to the rapid antidepressant effect of ketamine [19 ]. These findings are supported by those
of our study. We observed the antidepressant effects of low-dose ketamine infusion
only in patients without LGI. Therefore, inflammation-related depressive symptoms
may be resistant to the antidepressant effects of ketamine.
Our finding pertaining to the absence of the antidepressant effect in patients with
LGI may partially echo a clinical debate on whether anti-inflammatory drugs,
including nonsteroidal anti-inflammatory drugs, proinflammatory cytokine inhibitors,
and minocycline exert an adequately strong antidepressant effect against depression,
particularly TRD and bipolar depression [20 ]
[21 ]
[22 ]. A meta-analysis involving 6,262
individuals – of whom 4,258 were participants of nonsteroidal anti-inflammatory drug
trials and 2,004 were participants of cytokine inhibitor trials – demonstrated that
anti-inflammatory agents ameliorated depressive symptoms (pooled standard mean
difference: −0.34) compared with the effects of placebos [21 ]. However, the studies by Hellmann-Regen
et al. [20 ]
[22 ] and Husain et al. [20 ]
[22 ] demonstrated that 200 mg/day minocycline had no antidepressant effect
in patients with TRD or bipolar depression. By contrast, Nettis et al. demonstrated
that minocycline had antidepressant effects in patients with TRD with CRP levels
of≥3 mg/L [23 ]. Further studies are needed
to explore the effects of anti-inflammatory agents against depression, particularly
TRD, and to clarify whether ketamine’s inflammatory effect mediates its
antidepressant effect.
In this study, the antisuicidal effect of ketamine – measured in terms of the score
on item 10 of the MADRS – was observed in patients regardless of LGI status.
However, the effect was more prominent in patients without LGI than in those with
LGI. The antisuicidal effect of ketamine is partially independent of its
antidepressant effect, which potentially highlights the shared yet distinct
pathomechanisms underlying depression and suicide [24 ]
[25 ]. Inflammation leads to the production of quinolinic acid and
kynurenic acid, which are an agonist and an antagonist, respectively, of the
N-methyl-D-aspartate receptor [26 ]
[27 ]. Erhardt et al. [26 ] examined cerebrospinal fluid (CSF)
samples from 64 medication-free individuals who had attempted suicide and 36 control
individuals; the researchers found that the level of quinolinic acid but not
kynurenic acid was substantially elevated in the CSF of those who had attempted
suicide. Furthermore, levels of quinolinic acid and IL-6 in CSF were positively
correlated with total scores on the Suicide Intent Scale. This finding is consistent
with our finding that suggests that ketamine exerts antisuicidal effects in patients
with TRD regardless of LGI status.
In this study, the rate of response (on day 2) to the antidepressant effect of
ketamine was similar between the non-LGI group (HDRS: 41.9%; MADRS: 39.5%) and LGI
group (HDRS: 45.5%; MADRS: 42.4%); however, the rate of response (on day 2) to
placebo infusion was higher in the LGI group (HDRS: 30.8%; MADRS: 15.4%) than in the
non-LGI group (HDRS: 8.6%; MADRS: 5.7%). This difference explains why the
antidepressant effect of ketamine was noted only in the non-LGI group. To the best
of our knowledge, our study is the first to report this finding. The association
between inflammation and the placebo effect remains a mystery in neuroscience [28 ]
[29 ]
[30 ]. Kokkotou et al.
demonstrated a stronger placebo effect (symptom relief) in patients with irritable
bowel syndrome who had high levels of osteoprotegerin and TWEAK – two cytokines
belonging to the tumor necrosis factor superfamily – than in those who had low
levels of these cytokines [29 ]. Evans et
al. and Kokkotou et al. [29 ]
[30 ] proposed a counterintuitive hypothesis
suggesting that patients exhibiting a stronger immune response were more likely to
benefit from placebo treatment. In fact, despite the non-significance in the
distribution of infusion drugs (ketamine, midazolam, and normal saline) in our
study, more percentage of patients were allocated to midazolam than normal saline
(genuine placebo) in the LGI vs. non-LGI group, and more positive effects in the LGI
group may reflect this point since midazolam may have some antidepressant effects
[9 ]. Further randomized,
placebo-controlled trials investigating the effects of low-dose ketamine in patients
with TRD and LGI are needed to validate our findings.
Our study has several limitations. First, the study was a post hoc analysis of three
of our randomized placebo (normal saline or midazolam)-controlled and open-label
clinical trials. Combining data from clinical trials with different study designs
might have introduced a bias in our findings, necessitating further randomized,
large-scale, placebo-control trials to clarify the complex association of LGI with
the antidepressant and antisuicidal effects of ketamine. Second, our clinical trials
involved add-on ketamine therapy because the medications used by patients with TRD
were not discontinued during the trials. This is because an add-on design is
ethically appropriate for patients with TRD and offers real-life data. Third, CRP
levels may not be stable within the same individual. Further studies with repeated
CRP assessments may be required to define an LGI state in patients. Fourth, as
previously mentioned, the LGI prevalence was approximately 27% based on the CRP
level≥3 mg/L threshold [1 ]. Osimo et al.
further reported that the sample source (inpatient, outpatient, or
population-based), participant age, BMI, ethnicity, or antidepressant treatment did
not correlate with this LGI prevalence [1 ]. The definition of LGI based on CRP level≥3 mg/L may be arbitrary, but it
was reliable between our study and the meta-analysis study. Finally, despite the
exclusion of patients with current infectious diseases and other immunological
conditions from our clinical trials, we discovered that six patients had CRP levels
exceeding 10 mg/L. They did not exhibit any infectious symptoms or signs during the
clinical trials. After excluding those six patients, the analyses revealed
consistent findings.
In conclusion, among patients with TRD, only those without LGI respond to low-dose
ketamine infusion. However, further investigation is necessary to determine if
patients with TRD and LGI truly did not respond to low-dose ketamine, as our
analysis revealed a significant placebo effect in this group. In addition, ketamine
appears to exert antisuicidal effects in patients with TRD regardless of LGI status.
Further studies should be conducted to clarify the association between ketamine’s
anti-inflammatory effect and its antidepressant and antisuicidal effects.
Contributions
MHC and TPS designed the study. MHC, TPS, WCL, CTL, and HJW performed the clinical
trials; MHC analyzed the data and drafted the manuscript; TPS, CTL, WCL, YMB, SJT,
WCM, and PCT enrolled the candidate patients and performed the literature reviews.
All authors reviewed the final manuscript and agreed to its publication.