Abstract
OBJECTIVE
To describe the long-term outcomes, overall survival, progression-free survival, and prognostic factors in dogs with necrotizing encephalitis (NE).
ANIMALS
37 client-owned dogs clinically diagnosed with NE.
METHODS
All dogs underwent MRI and CSF analysis. Cox proportional hazards regression was used to examine factors related to the risk of relapse and death, including signalment, history, diagnostic investigation results, and treatments before the first relapse.
RESULTS
The medians of the overall and progression-free survival times were 639 days (IQR, 342 to 1,482 days) and 233 days (IQR, 111 to 775 days), respectively. Overall survival was highly correlated with progression-free survival. Four dogs (11%) died or were euthanized within 3 months of diagnosis. Relapse within 6 months was associated with a shorter overall survival. However, no prognostic factors for overall survival were found. The category of patients with presenting clinical signs that lasted 29 days to 6 months (OR, 3.26; 95% CI, 1.35 to 7.90) was associated with a higher risk of relapse. Seizures were presented in 75.7% of dogs, with a recurrence rate of 100%.
CLINICAL RELEVANCE
This report provides comprehensive follow-up information for dogs with NE, revealing a fair prognosis and low early mortality rate. Seizure is a very common clinical sign with a high recurrence rate.
Introduction
Necrotizing encephalitis (NE) encompasses the ante- and postmortem diagnoses of necrotizing meningoencephalitis (NME) and necrotizing leukoencephalitis (NLE) due to the significant overlap in clinical signs, signalment, and neuropathology observed in these conditions.1,2 Along with granulomatous meningoencephalomyelitis (GME), NE is included in the umbrella term meningoencephalomyelitis of unknown origin (MUO) for clinically diagnosed noninfectious inflammatory diseases of the CNS. Necrotizing encephalitis appears histopathologically distinct from GME because of the characteristic necrotic lesions in the cerebral white or gray matter.3,4 It has been reported that necrotic changes are usually absent in acute NME, and mild malacic lesions have been observed in a few cases of GME.3,4 Concomitant NME/NLE and GME in 4 dogs have been reported, suggesting these subtypes may overlap clinically and histopathologically in some cases.5 However, the presence of necrosis and cavitary lesions on MRI may help direct a presumptive antemortem diagnosis of NE.1
In the literature, most patients with NME and NLE have been noted to succumb to the disease within weeks to months.6–11 Only a few patients survived for 1 year with combined immunosuppressive therapies.12,13 The outcome data may be biased towards more severe cases because a histopathological diagnosis was required for the inclusion criteria in those reports. However, the data could affect owners’ and clinicians’ decision-making in evidence-based medicine. It has been proposed that dogs with MRI features compatible with acute and chronic NE should be excluded while recruiting MUO cases for clinical trials.14 The lack of studies focusing on long-term outcomes in clinically diagnosed NE cases impedes the discussion between owners and clinicians when applying evidence built on MUO studies.
In this retrospective study, we aimed to understand the prognostic factors and long-term outcomes of dogs with necrotic lesions on MRI that presented with typical characteristics of NE based on published literature. Overall survival (OS) was the primary end point used for evaluating prognostic indicators, along with a secondary end point of progression-free survival (PFS) before the first relapse. In veterinary medicine, confounding factor of the owner’s decision heavily affected the OS. It is conceivable that early relapse could influence the owner’s perception of the treatment’s feasibility and the patient’s quality of life. We hypothesized that individual PFS would highly correlate with OS. Evaluating prognostic indicators at this point may help identify patients with a higher risk of relapse and provide evidence for more aggressive therapy.
Methods
Study design and inclusion criteria
This was a single-center retrospective cohort study. Ethical approval by the committee of the National Taiwan University Veterinary Hospital (NTUVH112004) was obtained. All dog owners provided informed consent for using the clinical data in this study. All clinical examinations were performed by 2 board-certified neurologists or 1 neurology resident. Medical records from August 2007 to December 2020 were reviewed to identify dogs diagnosed with presumptive NE that had undergone treatment at our institution. Dogs were included if they (1) were older than 6 months at the onset of neurologic signs; (2) were of a breed reported to be affected with NME and NLE in the literature, including Chihuahuas9, French Bulldogs8, Maltese6, Yorkshire Terriers7, Pugs10, Coton de Tulears11, Papillons11, Shih Tzus11, Brussels Griffons11, Pekingese15, Pomeranians16, and Poodles17; (3) had undergone brain MRI demonstrating asymmetric intra-axial lesions that appeared iso- to hypointense on T1-weighted (T1W) and hyperintense on T2-weighted (T2W) sequences, with ≥ 1 lesion suggestive of a necrotic/cyst-like lesion characterized by being hyperintense on T2W and hypointense on T1W and T2-FLAIR sequences (Figure 1);1,14,18,19 (4) had undergone complete CSF analysis including total nucleated cell count (TNCC), total protein (TP), and cytology; (5) received negative test results for identifiable geographical infectious diseases where possible by PCR analysis on CSF and/or serology tests; (6) had medical records containing the date at the onset of neurologic signs; and (7) returned for follow-ups in our hospital at least until the first relapse or at least 6 months following diagnosis. Cases were excluded for any of the following reasons: (1) incomplete medical records for retrieving required information, (2) incomplete information regarding the neurological examination results in the follow-up visits before the first relapse, (3) a normal CSF analysis result, and (4) CSF exhibiting eosinophilic or neutrophilic pleocytosis based on the guideline of MUO diagnosis14 to rule out eosinophilic encephalitis and eliminate the possibility of including infectious encephalitis.
Transverse MRI images with T2-wighted (T2W) sequences (A), T1-weighted (T1W) sequences (B), and T2-FLAIR sequences (C). The inclusion criteria of brain MRI were dogs demonstrating intra-axial T1W iso- to hypointense, T2W and T2-FLAIR hyperintense lesions (arrowheads), with ≥ 1 lesion suggestive of a necrotic/cyst-like lesion characterized by being hyperintense on T2W and hypointense on T1W and T2-FLAIR images (arrows). R = Right. L = Left.
Citation: Journal of the American Veterinary Medical Association 262, 11; 10.2460/javma.24.03.0222
Diagnostic testing
An MRI was performed on all dogs under general anesthesia using either a 0.2-T magnet (Vet-MR; Esaote SpA) before 2019 or 1.5-T magnet (Ingenia S; Koninklijke Philips NV) after 2019. The minimum images required for evaluation included sagittal and transverse T2W, transverse FLAIR, and transverse T1W images before and after administration of gadolinium-based contrast medium. Data on the following MRI abnormalities were collected: (1) lesion location (ie, forebrain, cerebellum, and brainstem); (2) features indicating elevated intracranial pressure, including mass effect, effacement of the cerebral sulci, midline shift, transtentorial herniation, and foramen magnum herniation;20 and (3) the presence of moderate-to-marked contrast enhancement.
Cerebrospinal fluid from the cerebellomedullary or lumbar cisterns was collected immediately after MRI examination and analyzed within 30 minutes of collection. The results were categorized as normal or abnormal. Abnormal CSF was defined as having a TNCC > 5 cells/mm3, TP concentration > 25 mg/dL from the cerebellomedullary cistern or > 40 mg/dL from the lumbar area, or both.21 The cytologic classification of CSF was described as mononuclear or lymphocytic pleocytosis when > 70% of cells were mononuclear cells or lymphocytes, respectively. Neutrophilic or eosinophilic pleocytosis was observed when > 25% of neutrophils or > 1% of eosinophils were present, respectively. Mixed pleocytosis was considered when there was no clear predominance of a single-cell type.21
Data collection
The following data were retrieved from the medical records of each dog: signalment, duration of presenting clinical signs before admission, treatment before admission, CSF analysis results, MRI findings, date of diagnosis, OS before death or last follow-up, duration of PFS before the first relapse, treatment administered before the first relapse, neurological examination findings on admission and in the follow-ups, seizure history, presence of cluster seizure and/or status epilepticus (SE), and cause of death.
The duration of presenting clinical signs before admission was divided into 5 categories: ≤ 7 days, 8 to 14 days, 15 to 28 days, 28 days to 6 months, and > 6 months. Information associated with seizure history included the following: (1) presence of seizures before admission; (2) recurrent unprovoked seizures after diagnosis, which are defined as seizures with no identifiable acute CNS insults or metabolic disturbances after 7 days of diagnosis;22–25 and (3) anticonvulsive drug (ACD) administration at the time of admission. Cluster seizures were defined as ≥ 2 seizures within a 24-hour period. Status epilepticus was defined as the following: (1) > 5 minutes of continuous epileptic seizures or (2) ≥ 2 discrete epileptic seizures, between which there was incomplete recovery of consciousness. The presence of cluster seizures and/or SE was grouped together for analysis because the owners were sometimes unable to provide clear descriptions of the events.
Relapse was defined as the following: after clinical stabilization after treatment initiation, (1) the recurrence of previously described neurologic signs or the development of new signs indicating intracranial disease based on the owners’ description and neurologic examination results or (2) a sudden increase of > 50% in seizure frequency from the patient’s last presentation. Information regarding the treatment administered before the first relapse included the following: (1) the corticosteroid dose at the time of relapse and (2) whether and which add-on immunosuppressive agent was used among cytosine arabinoside (CA; Cytosar; Pfizer Inc), cyclosporine (Sandimmun Neoral; Novartis AG), and mycophenolate mofetil (MMF; CellCept; F Hoffmann-La Roche AG).
Statistical analysis
Continuous data were presented as mean and SD for normally distributed variables and as median and IQR for variables with nonnormal distribution. Categorical data were presented as numbers and percentages. Univariable and multivariable Cox proportional hazards regression models were used to investigate the association between independent variables and dichotomous outcomes, including relapse (ie, PFS) and death (ie, OS) at follow-up. The continuous independent variables included age at presentation, CSF TP levels, and CSF TNCC levels. Dichotomous independent variables included the following: sex; the presence or absence of seizure, cluster seizure, and/or SE; the duration of the presenting clinical signs before admission as ≤ 7 days, 8 to 14 days, 15 to 28 days, 28 days to 6 months, and > 6 months; lesions in forebrain, cerebellum, or brainstem; transtentorial herniation; midline shift; cerebral sulci effacement; foramen magnum herniation; mass effect; moderate-to-marked contrast enhancement; and the first add-on immunosuppressive medication before relapse as cyclosporin, MMF, and CA. Significant independent variables in the univariable analysis were entered into the multivariable model. Furthermore, significant independent variables in the multivariable model were identified as factors associated with dichotomous outcomes. The Kaplan-Meier survival function was used to observe the survival trend of PFS and OS in all patients and associated factor groups. The log-rank test was used to compare group differences in the Kaplan-Meier survival results. Spearman rank correlation was used to test for associations between the first relapse and death, PFS and OS, CSF findings, and all other independent variables. Statistical significance was set at 2-tailed P < .05. All analyses were performed using SPSS Statistics, version 25 (IBM Corp).
Results
Population characteristics
Between August 2007 and December 2020, 91 patients with MRI features of NE were reviewed. Among them, 15 had incomplete CSF examinations and/or pathogen screening results. Ten cases had normal CSF examination results. Another 27 patients had insufficient medical records for history and treatment outcome analyses. Two cases were excluded due to continuous progression of clinical signs despite treatment and death on days 35 and 36 after diagnosis. A total of 37 patients were eligible for this study, and their clinical characteristics and diagnostic test findings were analyzed (Table 1). Twenty male and 17 female dogs were included. The average age was 3.00 ± 1.90 years. Six breeds were represented. The most common pedigree breed was Maltese (n = 15), followed by Chihuahua (9), Yorkshire Terrier (5), French Bulldog (3), Pug (4), and Pomeranian (1).
Summary of clinical characteristics, MRI abnormalities, and CSF analysis findings in 37 dogs with presumptive necrotizing encephalitis. For MRI abnormalities, some cases had multiple locations and/or features.
| n | Percentage | |
|---|---|---|
| Age (y) | ||
| 0–4 | 26 | 70.3 |
| 5–8 | 10 | 27.0 |
| > 8 | 1 | 2.7 |
| Sex | ||
| Male | 20 | 54.0 |
| Female | 17 | 46.0 |
| Seizure history | ||
| Cases with seizures | 28 | 75.7 |
| Having the first seizure before presentation | 26 | 70.3 |
| Having the first seizure after presentation | 2 | 5.4 |
| Recurrent unprovoked seizure | 28 | 75.7 |
| Cluster seizure/SE | 18 | 48.6 |
| Duration of presenting signs | ||
| ≤ 7 d | 11 | 29.8 |
| 8 to 14 d | 3 | 6.4 |
| 15 to 28 d | 3 | 6.4 |
| 29 d to 6 mo | 11 | 23.4 |
| > 6 mo | 9 | 34.0 |
| CSF analysis | ||
| Elevated TP | 11 | 29.7 |
| Elevated TNCC | 2 | 5.4 |
| Elevated TP and TNCC | 24 | 64.9 |
| MRI abnormalities | ||
| Location | ||
| Forebrain | 27 | 73.0 |
| Brainstem | 14 | 37.8 |
| Cerebellum | 3 | 8.1 |
| MRI features of HICP | ||
| Transtentorial herniation | 3 | 8.1 |
| Midline shift | 5 | 13.5 |
| Cerebral sulci effacement | 13 | 2.7 |
| Foramen magnum herniation | 10 | 27.0 |
| Mass effect | 16 | 43.2 |
| Contrast enhancement | 16 | 43.2 |
| Treatment before the first relapse | ||
| Corticosteroid and cyclosporin | 13 | 35.1 |
| Corticosteroid and MMF | 7 | 18.9 |
| Corticosteroid and CA | 11 | 29.7 |
| MMF only | 1 | 2.7 |
| Corticosteroid only | ||
| Immunosuppressive dose | 3 | 8.1 |
| Anti-inflammatory dose | 1 | 2.7 |
| ACD only | 1 | 2.7 |
ACD = Anticonvulsive drug. CA = Cytosine arabinoside. HICP = High intracranial pressure. MMF = Mycophenolate mofetil. SE = Status epilepticus. TNCC = Total nucleated cell count. TP = Total protein.
The duration of presenting signs was categorized into 5 groups: ≤ 7 days (n = 11), between 8 and 14 days (3), between 15 and 28 days (3), between 29 days and 6 months (11), and > 6 months (9). At presentation, seizures were one of the chief complaints in 26 patients (70.3%). The distribution of seizure cases and whether the patient was on ACD before admission in each group are illustrated in Figure 2. For patients with a history of clinical signs lasting longer than 1 month, the prevalence of anticonvulsive treatment was higher (8/10 [80%]) in the beyond-6-months group compared to 2 of 5 (40%) in the 29-days-to-6-months group.
The distribution of seizure history and anticonvulsant drugs (ACD) in relation to the duration of clinical signs before admission. Dogs in each duration group were divided into 3 subgroups: no seizure (light gray), positive seizure history but not on ACD (dark gray), and positive seizure history and on ACD (black). D = Days. M = Months.
Citation: Journal of the American Veterinary Medical Association 262, 11; 10.2460/javma.24.03.0222
Diagnostic testing
The forebrain was the most common location for MRI lesions, accounting for 61.4% (27/44) of lesions (Table 1). Twelve patients had > 1 MRI feature suggesting possible elevated intracranial pressure, and 43.2% (16/37) of patients had moderate-to-marked contrast enhancement.
The CSF analysis revealed albuminocytological dissociation in 29.7% (11/37) of patients, elevated TNCC in 5.4% (2/37) of patients, and elevated TP and TNCC in 64.9% (24/37) of patients.
Treatment
Up to the time of the first relapse, 35 dogs had received immunosuppressive treatment, 1 dog had received only an anti-inflammatory dose of prednisolone, and 1 dog had received only anticonvulsive medication. Among the dogs undergoing immunosuppressive treatment, 3 dogs were treated only with prednisolone, 1 dog received MMF as the sole immunosuppressive agent, and 31 dogs received an immunosuppressive dose of prednisolone in combination with one of the add-on immunosuppressive agents. The initial protocols for add-on immunosuppressive agents were as follows:
Cyclosporine was given at 5 to 6 mg/kg, every 12 hours, in 13 dogs.
MMF was given at 10 mg/kg, every 12 hours, in 3 dogs and 20 mg/kg, every 24 hours, in 4 dogs.
CA was administered at 50 mg/m2, every 12 hours for 2 days, through SC injection in 11 dogs.
Among the 34 dogs in which prednisolone was given as part of immunosuppressive therapy, the initial dose was 2 mg/kg every 24 hours. The tapering protocol at our institution typically involved a 12.5% to 25% reduction of the dose every 3 to 4 weeks, as long as the dog’s neurologic status and blood results remained stable and no severe adverse reactions developed.
Outcome
Survival and relapse
The median survival time for all 37 dogs in this study was 639 days (range, 24 to 2,352 days; IQR, 342 to 1,482 days). The median PFS of all patients was 233 days (range, 24 to 1,590 days; IQR, 111 to 775 days). There was a significant correlation between PFS and OS (P < .001). In the dichotomous analysis, the time of the first relapse and death also showed a significant correlation (P < .001). The median dose of prednisolone at which dogs showed signs of the first relapse was 0.75 mg/kg, every 24 hours, (IQR, 0.28 to 1.23) with a median time of 198 days (IQR, 81 to 267).
The survival trends between dogs that had no relapse and first relapse within and after 6 months is illustrated in Figure 3. The survival trends of dogs with no relapse and those with a first relapse after 6 months were similar. A significant difference (P < .001) was observed in the Kaplan-Meier survival function between dogs with and without relapse within 6 months (Figure 4). Twenty-three patients (62.2%) died during the follow-up. The median duration of the follow-up was 556 days (IQR, 266 to 1,007). The median survival time of all deceased patients was 395 days (IQR, 112 to 749). Only 4 patients died within 3 months of diagnosis.
A Sankey diagram illustrating the outcome in 37 dogs diagnosed with presumptive necrotizing encephalitis in terms of survival divided by the presence of relapse and relapse occurring 6 months (m) before or after diagnosis. Outcomes are written horizontally in each node (survived, dead, or lost to follow-up), with the node height representing the number of dogs affected. Numbers of dogs for each outcome are written vertically in each node.
Citation: Journal of the American Veterinary Medical Association 262, 11; 10.2460/javma.24.03.0222
Kaplan-Meier curves showing overall survival in dogs with presumptive necrotizing encephalitis. Dogs experiencing relapse within 6 months (green) had significantly shorter survival than dogs without relapses within 6 months (blue). Tick marks indicate censored cases. Cum = Cumulative.
Citation: Journal of the American Veterinary Medical Association 262, 11; 10.2460/javma.24.03.0222
Cause of death
Among the 23 deceased dogs, 11 dogs were euthanized and 12 dogs were reported dead by the owners. Seizure frequency and/or deteriorating neurological conditions were recorded in 19 dogs before euthanasia (10/11) or death (9/12). Comorbidities before euthanasia, including aspiration pneumonia, anemia, and gastrointestinal bleeding, were documented in 4 dogs. Sudden unexpected death was recorded in 2 other dogs, both Pugs. The other 2 Pugs died during or immediately after a seizure witnessed by the owner.
Prognostic factors and survival analysis
Since only 1 significant variable was found in the univariable analysis for PFS using the Cox proportional hazards regression model (Table 2), the multivariable models were omitted for both the PFS and OS results. Dogs with 29 days’ to 6 months’ duration of presenting signs (OR, 3.26; 95% CI, 1.35 to 7.90; P = .009) were associated with a higher risk of relapse. A marginal significance of using cyclosporin as the first add-on immunosuppressant in addition to corticosteroids before the first relapse (OR, 2.23; 95% CI, 0.99 to 5.01; P = .053) was noticed in the univariable analysis for PFS. Survival estimates of these 2 factors on PFS were analyzed using Kaplan-Meier survival curves. Grouped according to 29 days’ to 6 months’ duration of presenting signs, the median PFS in dogs with and without this factor was 190 and 426 days, respectively (P = .006). When grouped according to cyclosporine as the first add-on immunosuppressant before the first relapse, the median PFS of patients with and without cyclosporine was 203 and 250 days, respectively (P = .048).
Cox proportional hazards models of independent variables to relapse (progression-free survival) and death outcome (overall survival).
| Variables | Progression-free survival OR (95% CI) | P value | Overall survival OR (95% CI) | P value |
|---|---|---|---|---|
| Age | 0.99 (0.81–1.21) | .908 | 1.13 (0.90–1.42) | .299 |
| Sex | ||||
| Male | ref | – | ref | – |
| Female | 0.89 (0.42–1.87) | .755 | 1.17 (0.50–2.71) | .720 |
| Seizure | 0.53 (0.23–1.20) | .126 | 1.17 (0.39–3.50) | .784 |
| Cluster seizure/SE | 1.28 (0.61–2.72) | .517 | 1.23 (0.53–2.86) | .634 |
| Duration of presenting signs | ||||
| ≤ 7 d | 0.61 (0.27–1.39) | .240 | 1.02 (0.41–2.54) | .969 |
| 8 to 14 d | 1.30 (0.39–4.38) | .670 | 1.02 (0.24–4.41) | .974 |
| 15 to 28 d | 0.63 (0.30–1.35) | .233 | 0.31 (0.04–2.45) | .268 |
| 29 d to 6 mo | 3.26 (1.35–7.90) | .009 | 1.32 (0.55–3.16) | .531 |
| > 6 mo | 0.62 (0.23–1.68) | .348 | 1.07 (0.39–2.90) | .897 |
| CSF analysis | ||||
| TP | 1.00 (0.99–1.01) | .987 | 1.00 (0.99–1.01) | .872 |
| TNCC | 1.00 (1.00–1.01) | .581 | 1.00 (1.00–1.00) | .772 |
| MRI abnormalities | ||||
| Location | ||||
| Forebrain | 1.19 (0.52–2.72) | .679 | 1.67 (0.65–4.27) | .283 |
| Brainstem | 1.90 (0.87–4.14) | .106 | 1.22 (0.49–3.03) | .666 |
| Cerebellum | 1.50 (0.34–6.49) | .592 | 0.84 (0.18–3.86) | .824 |
| MRI features of HICP | ||||
| Transtentorial herniation | 0.70 (0.17–2.97) | .630 | 0.69 (0.09–5.17) | .718 |
| Midline shift | 2.14 (0.72–6.40) | .174 | 2.31 (0.75–7.16) | .145 |
| Cerebral sulci effacement | 1.17 (0.53–2.55) | .703 | 1.48 (0.60–3.63) | .391 |
| Foramen magnum herniation | 0.46 (0.18–1.16) | .101 | 0.46 (0.17–1.26) | .129 |
| Mass effect | 1.76 (0.83–3.72) | .139 | 1.21 (0.53–2.75) | .652 |
| Contrast enhancement | 1.37 (0.65–2.91) | .409 | 1.28 (0.56–2.91) | .559 |
| Immunosuppressants | ||||
| Cyclosporin | 2.23 (0.99–5.01) | .053 | 1.97 (0.83–4.64) | .122 |
| MMF | 0.58 (0.22–1.55) | .279 | 0.50 (0.17–1.71) | .268 |
| CA | 0.98 (0.45–2.13) | .961 | 0.67 (0.27–1.67) | .387 |
Ref = the reference group in the variable of sex.
Discussion
This study focused on dogs with suspected necrotic lesions on MRI images presenting with typical characteristics of NE, as noted in published literature.2,18 Although histopathological findings have been reported to be relatively specific for NME and NLE, these 2 subtypes of NE appear to have considerable overlap in breed association and lesion distribution.1,2 The variable degrees of necrosis and lesion topography between NME and NLE may reflect minor genotypic differences within and among breeds.11 The median OS of the dogs in this study was 639 days. In previous reports,6–11 dogs with NME and NLE usually succumbed to the disease within weeks to months. Improved survival, ranging from 177 days to 306 days, can be achieved with a combination of immunosuppressive drugs and corticosteroids.10,12 Euthanasia is not widely accepted in our culture, and the experience of managing the disease has increased over the last decade. A total of 83% of the dogs in this study received combined immunosuppressive therapy, which may have contributed to the longer survival observed in this report. In contrast to our study, previous reports6–11 on NME and NLE have been histopathologically confirmed, creating a possible bias towards more severe cases. We selected cases with necrotic and cavitary lesions on MRI and complete CSF analysis. However, this could have led to case selection for dogs in the subacute or chronic phases of NME, dogs with fewer concerns in terms of elevated intracranial pressure, or dogs with GME and necrotic lesions. Another factor that could influence OS was that, since we included relapse time as a secondary end point, patients that were not stabilized following treatment before relapse were excluded from our study. Two patients were excluded for this reason and died 35 and 36 days after treatment initiation. There are 2 studies20,26 focusing on the short-term survival in MUO cases. The prospective study20 reported that the median survival time in all deceased dogs was 2 days, and 33% of these patients died within 3 days of diagnosis. The retrospective study26 on 116 dogs with MUO reported that 26% died within 7 days of diagnosis. Both studies included patients with intra-axial T2W hyperintense lesions, but dogs with T1W hypointense cyst-like lesions were excluded in one study,20 and the number of dogs with cyst-like lesions in the other study26 is not known. In contrast to these 2 studies, the median survival time in all the deceased dogs in our study was 395 days. Short-term mortality may be different in dogs with MUO with suspected necrotic/cyst-like lesions, forming a clinical portrait of a protracted and debilitating process.
Progression-free survival was used as the secondary end point in this study. This was significantly correlated with OS. A previous study20 on MUO determined that relapse is not associated with the outcome. The presence of hypointense cystic cortical lesions on T1W images was an exclusion criterion in that report, which may underlie the variances between NE and other forms of MUO. One reason for the addition of a secondary end point for prognostic factor evaluation is the relative uniformity of treatment protocols before progression. In addition, in our experience, owners are usually willing to try treatment rather than euthanasia after a diagnosis due to cultural beliefs. Versatile adjustments were initiated after the first relapse, including involving various dosage modification protocols, adding another immunosuppressive agent, and analgesics. In addition, 2 dogs had received acupuncture therapy in private practices. Thus, the length of postprogression survival was highly affected by random variations associated with patient heterogeneity, owners’ perception of euthanasia and quality of life, and the influence of subsequent therapy. Using a secondary end point, we identified a group of patients with a higher risk of relapse due to delayed treatment between 29 days and 6 months after the onset of clinical signs. Delayed treatment may reflect the severity of these cases, as the patients’ conditions may have been aggravated or could not have been stabilized in primary care clinics 1 month after the onset of clinical signs. Inadequate seizure control was also observed in this patient group. Among the 5 of 10 patients in this group that had a history of seizures before admission, only 2 of 5 (40%) had received anticonvulsive drugs. Compared to the 10 patients with clinical signs for > 6 months, all patients had seizure histories and 8 of 10 (80%) of them had received ACD before admission. Due to the high variation in the dosages and medications they received, this factor was not listed for prognostic evaluation. In addition, the lack of a precise seizure log made it unknown whether seizure severity affected the decision to use ACD in these 2 groups of patients. Studies27,28 have shown that recurrent seizures damage the hippocampus and blood-brain barrier due to seizure-induced brain inflammation and excitotoxicity. Interestingly, the prevalence of recurrent unprovoked seizures in our study varied from that which is reported in published literature on MUO. A retrospective study25 exploring the prevalence of postencephalitic epilepsy in 61 dogs with MUO distinguished acute symptomatic seizures from late unprovoked seizures. In 29 dogs with seizures preceding or occurring within 1 week of MUO diagnosis, only 11 dogs developed postencephalitic recurrent unprovoked seizures in that report. In contrast, all patients with seizures in our study had postencephalitic recurrent unprovoked seizures. To our knowledge, no studies have compared the differences in epileptogenesis among the subtypes of canine noninfectious meningoencephalitis. In humans, the 4 clinical factors with the highest risk of seizure recurrence include EEG with epileptiform abnormalities, prior brain insult such as stroke or head trauma, significant de novo brain-imaging abnormality, and nocturnal seizure.29 Static brain lesions caused by traumatic brain injury, stroke, or infection are associated with an enduring predisposition to seizures.30 Cortical involvement and multifocal areas of ischemia are the main risk factors for poststroke seizures.31 A similar condition could exist in NE, especially in NME, in which dogs have static necrotic and cavitary lesions involving multiple cerebral cortices. In humans, although anticonvulsive treatment after a second unprovoked seizure does not affect the prognosis of epilepsy development, early anticonvulsive treatment is likely to reduce the risk of a second seizure. After a patient experiences a second seizure, the risk of subsequent seizures increases by almost 60% by the first year.29 Early medical intervention for seizures is worth emphasizing in dogs with compatible signalments of NE.
In dogs, dosages of 0.5 to 1.0 mg/kg of prednisolone, PO, every 24 hours, are considered to have anti-inflammatory effects.32 Our study found that the median dose of prednisolone at which dogs showed signs of first relapse was 0.75 mg/kg, every 24 hours. This finding may imply that some patients require robust or long-term synergistic immunosuppressive protocols. At our institution, immunosuppressive therapy with prednisolone was started at a dose of 2 mg/kg, every 24 hours, typically followed by a 12.5% to 25% reduction in the dose every 3 to 4 weeks if the dog’s neurological status and bloodwork remained stable and no severe adverse reactions developed. However, the tapering protocol was not strictly standardized and was influenced by the individual patient’s condition, the owner’s preference, and the clinician’s decision. The lack of standardized treatment protocols, mixed populations of suspected NME and NLE, and the small number of patients in this study necessitate a cautious interpretation of this finding. Implementing therapeutic biomarker monitoring such as neurofilament light chain with neurologic examination can help evaluate the treatment response and individualize therapeutic interventions in dogs with MUO.33
In the literature, NME in Pugs is considered an early-onset, rapidly progressing disease with a short median survival time of 23 to 93 days. The median age at the onset of clinical signs in Pugs is approximately 1.5 years.10,34,35 The Pugs in our study were older, with a median age of 5.5 years and a longer median survival of 464 days. Potential biases of this finding include the small sample size and the absence of histopathological confirmation. It is important to note that the finding could be influenced by the fact that the Pugs in this study may have other CNS disorders or that NME patients in confirmed cases may have worse outcomes in the literature. However, it is also possible that the genotypes, individual susceptibilities, and underlying triggers in Taiwan are different from those in other geographic areas. A recent study36 showed that some Pugs with risk genotypes could have an abnormal neurological examination of multifocal spinal hyperesthesia that is unconventional in NME and MRI findings indicating inflammatory brain diseases while remaining asymptomatic to the owners. Our patients may have had similar subclinical manifestations before presentation, or there was a less aggressive and chronic form of NME, considering their much longer survival time after exhibiting conventional NME signs. None of the Pugs in this study underwent germline risk assessment. It would be worthwhile for future studies to obtain more extensive risk genotype and systemic inflammatory profiles in Pugs in Taiwan.
In addition to those mentioned above, including the case selection, mixture of suspected NME and NLE cases, lack of a control group and standardized treatment protocols, possible divergent views on euthanasia between Asian and Western cultures, nonanonymized observers, and the small sample size, this study has several other limitations. First, histopathological confirmation of the diagnosis was not possible in all cases. Only 1 case in this report had a histopathology examination and confirmed diagnosis. Although strict inclusion criteria were applied, it is possible that some dogs with other CNS disorders, such as GME, neoplasia, CNS infectious diseases, or chronic cerebrovascular lesions, were included. Cyst-like lesions are not listed as an MRI feature of GME.37 Neoplasia and vascular events are the 2 other common etiological categories of acquired cysts. However, the specificity of differentiating NE from neoplastic and cerebrovascular diseases was 92.9% in 1 study.38 Nevertheless, we understand that the gold standard for diagnosing those intracranial diseases remains histopathology.1 Concomitant NE and granulomatous meningoencephalitis have also been reported.39 Second, a change in MRI equipment at our institution occurred during the case-selection time frame. The installation of a high-field MRI could lead to a better appreciation of subtle abnormalities, such as contrast enhancement. However, the bias was partially diminished by collecting only moderate-to-marked MRI findings for the prognostic evaluation.
In conclusion, in dogs with presumptive NE, relapse within 6 months was associated with a shorter OS time. The timing of medical intervention is an important factor that influences relapse. The seizures in our dogs were all recurrent, supporting the rationale for starting anticonvulsive treatment early in dogs with compatible NE signalment. We also confirmed a strong correlation between PFS and OS; however, the surrogacy of PFS for OS needs to be verified in clinical trials. Necrotizing encephalitis is a subgroup of MUO; nevertheless, variances in clinical presentation, progression, and response to treatment may increase the heterogeneity of the study groups in research on MUO cases.
Acknowledgments
The authors gratefully acknowledge Dr. Chun-Sheng (Jason) Lee for his involvement with patient management and Jhih-Jyun Yang for statistics consultation.
Disclosures
The authors have nothing to disclose. No AI-assisted technologies were used in the generation of this manuscript.
Funding
This research was funded by the National Taiwan University Veterinary Hospital (C11204). The funding agencies had no role in the study design, data acquisition, analysis, or interpretation of this work.
ORCID
Y.-P. Chang https://orcid.org/0000-0002-5123-7062
References
- 1.↑
Talarico LR, Schatzberg SJ. Idiopathic granulomatous and necrotising inflammatory disorders of the canine central nervous system: a review and future perspectives. J Small Anim Pract. 2010;51(3):138-149. doi:10.1111/j.1748-5827.2009.00823.x
- 2.↑
Coates JR, Jeffery ND. Perspectives on meningoencephalomyelitis of unknown origin. Vet Clin North Am Small Anim Pract. 2014;44(6):1157-1185. doi:10.1016/j.cvsm.2014.07.009
- 3.↑
Park ES, Uchida K, Nakayama H. Comprehensive immunohistochemical studies on canine necrotizing meningoencephalitis (NME), necrotizing leukoencephalitis (NLE), and granulomatous meningoencephalomyelitis (GME). Vet Pathol. 2012;49(4):682-692. doi:10.1177/0300985811429311
- 4.↑
Uchida K, Park E, Tsuboi M, Chambers JK, Nakayama H. Pathological and immunological features of canine necrotising meningoencephalitis and granulomatous meningoencephalitis. Vet J. 2016;213:72-77. doi:10.1016/j.tvjl.2016.05.002
- 5.↑
Nessler JN, Oevermann A, Schawacht M, et al. Concomitant necrotizing encephalitis and granulomatous meningoencephalitis in four toy breed dogs. Front Vet Sci. 2022;9:957285. doi:10.3389/fvets.2022.957285
- 6.↑
Stalis IH, Chadwick B, Dayrell-Hart B, Summers BA, Van Winkle TJ. Necrotizing meningoencephalitis of Maltese dogs. Vet Pathol. 1995;32(3):230-235. doi:10.1177/030098589503200303
- 7.↑
Kuwamura M, Adachi T, Yamate J, Kotani T, Ohashi F, Summers BA. Necrotising encephalitis in the Yorkshire Terrier: a case report and literature review. J Small Anim Pract. 2002;43(10):459-463. doi:10.1111/j.1748-5827.2002.tb00014.x
- 8.↑
Timmann D, Konar M, Howard J, Vandevelde M. Necrotising encephalitis in a French Bulldog. J Small Anim Pract. 2007;48(6):339-342. doi:10.1111/j.1748-5827.2006.00239.x
- 9.↑
Higgins RJ, Dickinson PJ, Kube SA, et al. Necrotizing meningoencephalitis in five Chihuahua dogs. Vet Pathol. 2008;45(3):336-346. doi:10.1354/vp.45-3-336
- 10.↑
Levine JM, Fosgate GT, Porter B, Schatzberg SJ, Greer K. Epidemiology of necrotizing meningoencephalitis in Pug dogs. J Vet Intern Med. 2008;22(4):961-968. doi:10.1111/j.1939-1676.2008.0137.x
- 11.↑
Cooper JJ, Schatzberg SJ, Vernau KM, et al. Necrotizing meningoencephalitis in atypical dog breeds: a case series and literature review. J Vet Intern Med. 2014;28(1):198-203. doi:10.1111/jvim.12233
- 12.↑
Jung DI, Kang BT, Park C, et al. A comparison of combination therapy (cyclosporine plus prednisolone) with sole prednisolone therapy in 7 dogs with necrotizing meningoencephalitis. J Vet Med Sci. 2007;69(12):1303-1306. doi:10.1292/jvms.69.1303
- 13.↑
Jung DI, Kim JW, Park HM. Long-term immunosuppressive therapy with cyclosporine plus prednisolone for necrotizing meningoencephalitis in a Pekingese dog. J Vet Med Sci. 2012;74(6):765-769. doi:10.1292/jvms.11-0468
- 14.↑
Granger N, Smith PM, Jeffery ND. Clinical findings and treatment of non-infectious meningoencephalomyelitis in dogs: a systematic review of 457 published cases from 1962 to 2008. Vet J. 2010;184(3):290-297. doi:10.1016/j.tvjl.2009.03.031
- 15.↑
Cantile C, Chianini F, Arispici M, Fatzer R. Necrotizing meningoencephalitis associated with cortical hippocampal hamartia in a Pekingese dog. Vet Pathol. 2001;38(1):119-122. doi:10.1354/vp.38-1-119
- 16.↑
Matsuki N, Takahashi M, Yaegashi M, Tamahara S, Ono K. Serial examinations of anti-GFAP autoantibodies in cerebrospinal fluids in canine necrotizing meningoencephalitis. J Vet Med Sci. 2009;71(1):99-100. doi:10.1292/jvms.71.99
- 17.↑
Song JH, Moon JH, Kim HS, et al. Acute progress of necrotizing meningoencephalitis in a dog; serial clinical observation, magnetic resonance imaging, and histopathological findings. J Vet Clin. 2016;33(5):290-294. doi:10.17555/jvc.2016.10.33.5.290
- 18.↑
Flegel T. Breed-specific magnetic resonance imaging characteristics of necrotizing encephalitis in dogs. Front Vet Sci. 2017;4:203. doi:10.3389/fvets.2017.00203
- 19.↑
Young BD, Levine JM, Fosgate GT, et al. Magnetic resonance imaging characteristics of necrotizing meningoencephalitis in Pug dogs. J Vet Intern Med. 2009;23(3):527-535. doi:10.1111/j.1939-1676.2009.0306.x
- 20.↑
Lowrie M, Smith PM, Garosi L. Meningoencephalitis of unknown origin: investigation of prognostic factors and outcome using a standard treatment protocol. Vet Rec. 2013;172(20):527. doi:10.1136/vr.101431
- 21.↑
De Lorenzi D, Mandara MT. The central nervous system. In: Raskin RE, Meyer DJ. Canine and Feline Cytology. 3rd ed. Elsevier Saunders; 2016:369-407.
- 22.↑
International League Against Epilepsy. Guidelines for epidemiologic studies on epilepsy: commission on epidemiology and prognosis, international league against epilepsy. Epilepsia. 1993;34(4):592-596. doi:10.1111/j.1528-1157.1993.tb00433.x
- 23.
Beghi E, Carpio A, Forsgren L, et al. Recommendation for a definition of acute symptomatic seizure. Epilepsia. 2010;51(4):671-675. doi:10.1111/j.1528-1167.2009.02285.x
- 24.
Forsgren L, Bucht G, Eriksson S, Bergmark L. Incidence and clinical characterization of unprovoked seizures in adults: a prospective population-based study. Epilepsia. 1996;37(3):224-229. doi:10.1111/j.1528-1157.1996.tb00017.x
- 25.↑
Kaczmarska A, José-López R, Czopowicz M, et al. Postencephalitic epilepsy in dogs with meningoencephalitis of unknown origin: clinical features, risk factors, and long-term outcome. J Vet Intern Med. 2020;34(2):808-820. doi:10.1111/jvim.15687
- 26.↑
Cornelis I, Volk HA, Van Ham L, De Decker S. Prognostic factors for 1-week survival in dogs diagnosed with meningoencephalitis of unknown aetiology. Vet J. 2016;214:91-95. doi:10.1016/j.tvjl.2016.05.008
- 27.↑
Lado FA, Laureta EC, Moshé SL. Seizure-induced hippocampal damage in the mature and immature brain. Epileptic Disord. 2002;4(2):83-97. doi:10.1684/j.1950-6945.2002.tb00479.x
- 28.↑
Librizzi L, Noè F, Vezzani A, de Curtis M, Ravizza T. Seizure-induced brain-borne inflammation sustains seizure recurrence and blood-brain barrier damage. Ann Neurol. 2012;72(1):82-90. doi:10.1002/ana.23567
- 29.↑
Rizvi S, Ladino LD, Hernandez-Ronquillo L, Téllez-Zenteno JF. Epidemiology of early stages of epilepsy: risk of seizure recurrence after a first seizure. Seizure. 2017;49:46-53. doi:10.1016/j.seizure.2017.02.006
- 30.↑
Hesdorffer DC, Benn EKT, Cascino GD, Hauser WA. Is a first acute symptomatic seizure epilepsy? Mortality and risk for recurrent seizure. Epilepsia. 2009;50(5):1102-1108. doi:10.1111/j.1528-1167.2008.01945.x
- 31.↑
Doria JW, Forgacs PB. Incidence, implications, and management of seizures following ischemic and hemorrhagic stroke. Curr Neurol Neurosci Rep. 2019;19(7):37. doi:10.1007/s11910-019-0957-4
- 32.↑
Pettersson H, Ekstrand C, Hillström A, Lilliehöök I. Effect of 1 mg/kg oral prednisolone on biochemical analytes in ten dogs: a cross-over study. Comp Clin Path. 2021;30(3):519-528. doi:10.1007/s00580-021-03246-9
- 33.↑
Windsor R, Stewart SD, Talboom J, et al. Leukocyte and cytokine variables in asymptomatic Pugs at genetic risk of necrotizing meningoencephalitis. J Vet Intern Med. 2021;35(6):2846-2852. doi:10.1111/jvim.16293
- 34.↑
Cordy DR, Holliday TA. A necrotizing meningoencephalitis of Pug dogs. Vet Pathol. 1989;26(3):191-194. doi:10.1177/030098588902600301
- 35.↑
Kobayashi Y, Ochiai K, Umemura T, Goto N, Ishida T, Itakura C. Necrotizing meningoencephalitis in Pug dogs in Japan. J Comp Pathol. 1994;110(2):129-136. doi:10.1016/S0021-9975(08)80184-3
- 36.↑
Windsor R, Stewart S, Schmidt J, et al. A potential early clinical phenotype of necrotizing meningoencephalitis in genetically at-risk Pug dogs. J Vet Intern Med. 2022;36(4):1382-1389. doi:10.1111/jvim.16444
- 38.↑
Wolff CA, Holmes SP, Young BD, et al. Magnetic resonance imaging for the differentiation of neoplastic, inflammatory, and cerebrovascular brain disease in dogs. J Vet Intern Med. 2012;26(3):589-597. doi:10.1111/j.1939-1676.2012.00899.x
- 39.↑
Wanamaker MW, Vernau KM, Taylor SL, Cissell DD, Abdelhafez YG, Zwingenberger AL. Classification of neoplastic and inflammatory brain disease using MRI texture analysis in 119 dogs. Vet Radiol Ultrasound. 2021;62(4):445-454. doi:10.1111/vru.12962
