Investigation of the therapeutic effects, predictors, and complications of long-term immunosuppressive therapy in dogs with precursor-targeted immune-mediated anemia

Abstract

Dogs with precursor-targeted immune-mediated anemia (PIMA) are commonly treated with immunosuppressive therapy, but information on predictors of treatment response and response time is limited. Therefore, we retrospectively investigated predictive factors that influenced the treatment response and duration required to observe a response in dogs with PIMA receiving continuous immunosuppressive therapies for more than 105 days. Of 50 client-owned dogs that developed PIMA, 27 were included in this study, of which 18 were responders and 9 were non-responders to immunosuppressive therapies. Sixteen of the 18 responders responded to treatment within 60 days and the remaining 2 responded at 93 and 126 days, respectively. We found that an erythroid-maturation ratio of <0.17 may be a useful predictor for treatment response. In addition, complications of immunosuppressive therapies were investigated further in 50 dogs. Pancreatitis (n=4) and pneumonia (3) occurred over the entire treatment period, and infections such as abscesses (3) tended to be more common in dogs on an extended period of immunosuppressive therapy. These findings may be helpful when planning for the initial treatment and may provide evidence for informed consent about potential comorbidities throughout the treatment course.

Dogs with persistent non-regenerative anemia show ineffective erythropoiesis in the bone marrow, which is characterized as erythroid hyperplasia or maturation arrest. Although the pathogenesis remains unclear, these cases are considered a type of immune-mediated hematologic disease because some cases showed favorable responses to immunosuppressive therapy, whereas other cases relapsed when the therapy was tapered [11,12,13]. Because of similarities with immune-mediated hemolytic anemia, this disease has been termed non-regenerative immune-mediated anemia (NRIMA) [13, 15, 18]. However, in the absence of the significant immune-mediated destruction of peripheral blood cells, precursor-targeted immune-mediated anemia (PIMA), proposed by Lucidi et al. in 2017, is now considered a more appropriate diagnostic term [8].

Previous reports have described the clinicopathologic characteristics of dogs diagnosed with NRIMA/PIMA and their outcomes following immunosuppressive therapy. Several retrospective studies indicated that the median time to treatment response was about 14–31 days [1, 8, 13, 18], but there was significant variation among individuals, with some dogs requiring as long as 105 days [8, 16, 18]. Although some recent studies also examined predictors for survival, these studies were limited by the inclusion of cases in which the duration of immunosuppressive therapy was too short to assess the treatment response [1, 18]. Therefore, it would be more appropriate to identify predictive factors influencing treatment responses only in dogs who have received immunosuppressive therapy for more than 105 days. Furthermore, in NRIMA/PIMA dogs, the prognosis is expected to be influenced by treatment response as well as complications associated with prolonged severe anemia or immunosuppressive therapy. However, the type, frequency, and timing of complications and adverse events are not known.

The purpose of this study was to investigate predictors of the therapeutic effect in PIMA dogs receiving immunosuppressive therapy for more than 105 days, which is considered a sufficient duration to determine treatment responses. In addition, complications or adverse events in PIMA dogs treated with immunosuppressive treatment, including those treated for less than 105 days, were summarized as evidence for informed consent to assist in the initiation of treatment.

MATERIALS AND METHODS

Case selection

First, client-owned dogs diagnosed with NRIMA or PIMA at Hokkaido University Veterinary Teaching Hospital between December 2012 and August 2020 were identified by electronic medical record review. Second, each dog was judged to ensure they met the diagnostic criteria of PIMA: (i) minimum 5-day history of severe non-regenerative anemia (hematocrit <20%) with an absolute reticulocyte count <60 × 10³/μL [13, 15]; (ii) bone marrow (BM) cytology and/or histopathology results available and ineffective erythropoiesis identified as erythroid hypercellularity or an increase in early-stage erythroid precursors with maturation arrest that could not be attributed to a pre-regenerative response or other disease process (e.g., myelodysplastic disease or drug reaction); and (iii) complete blood count, serum biochemistry profile, and thoracic and abdominal imaging results revealed no underlying cause for severe non-regenerative anemia. Dogs were included if they fulfilled all the following data: signalments; clinical signs noticed by the owner; treatment before referral; physical examination findings; laboratory findings including complete blood count (CBC) conducted within 24 hr of bone marrow extraction or before their latest blood transfusion, corrected reticulocyte % (CR%) [3], erythrocyte agglutination test; Coombs test; peripheral blood smear evaluation; treatment after referral; and treatment outcomes. Immunosuppressive therapy (e.g., more than 2 mg/kg of prednisolone with one or more immunosuppressive drugs) that was already started at the referral hospital was included in the treatment duration.

Assessment of response to immunosuppressive therapy

To evaluate the response of long-term immunosuppressive therapy, only dogs that had been treated for more than 105 days were included in this assessment. Dogs were considered to have responded to immunosuppressive therapy if their absolute reticulocyte count increased to ≥60,000 /μL or they had a spontaneous increase in the packed cell volume (PCV) of >5%. A sole increase in reticulocyte without an increase in PCV was not considered a treatment response. Dogs that failed to achieve the therapeutic response after a duration exceeding 105 days of immunosuppressive therapy were classified as non-responsive to immunosuppression. Relapse was defined as a worsening of non-regenerative anemia requiring the intensification of immunosuppressive therapy in responders.

BM cytology samples were evaluated by one clinician who was blinded to the case information, using BM smears obtained at the time of diagnosis. A total of 500 cells per slide were evaluated. The absence of polychromatophils and presence of rubriphagocytosis, and the macrophage percentage of all nucleated cells, were recorded. The myeloid-to-erythroid ratio, granulocyte-maturation ratio (GMR), and erythroid-maturation ratio (EMR) were investigated, as reported previously [14, 17]. GMR was calculated using the following formula: (myeloblasts + promyelocytes + myelocytes)/(metamyelocytes + band neutrophils + segmented neutrophils). EMR was calculated using the following formula: (rubriblasts + prorubricytes)/(rubricytes + metarubricytes). The reference intervals (RI) of GMR and EMR were set at 0.04–0.15 and 0.02–0.12, respectively, according to a previous study [17]. A pathological examination of the BM was performed by a pathologist to assess fibrosis and nucleated cell and erythroblast fullness. The extent of fibrosis (mild, moderate, or severe) was graded subjectively, as described previously [1].

Investigation of adverse events and complications

Problems other than anemia that occurred during the implementation of immunosuppressive therapy were investigated. These included complications associated with PIMA and side effects from immunosuppressive drugs. Side effects that improved with the tapering of medications such as polyuria and polydipsia or did not require a new therapeutic intervention for the problem that arose were excluded. Because it was not always easy to distinguish between complications and side effects, we used the term complications to describe both.

Statistical analysis

All analyzes were performed using commercial software (JMP Pro 16.0.0, SAS Institute Inc., Cary, NC, USA). Statistical significance was set as P<0.05. Signalments, CBC, and BM cytology results were analyzed and compared between responders and non-responders to immunosuppressive treatment. Continuous variables (age, body weight, PCV, reticulocytes, platelets, CR%, EMR, GMR, and macrophage ratio) were analyzed by one-sided Wilcoxon’s exact test. Sex, neutering status, existence or non-existence of polychromatophils and rubriphagocytosis in BM smears were analyzed by Fisher’s exact test. Logistic regression analysis was performed using continuous variables that showed significant differences in the comparison between the two groups. From the results obtained, a receiver operating characteristic curve drawn, and corresponding areas under the curve (AUC) were calculated to assess the predictive accuracies. The cutoff value of the continuous variable with the highest AUC was determined using the Youden Index.

RESULTS

Population characteristics

During the observation period, 50 cases presented to our clinic were diagnosed with PIMA based on bone marrow examination. The median age was 10 years (range: 4–14 years), and the dogs included 30 females (17 intact and 13 neutered) and 20 males (2 intact and 18 neutered). There were 10 different breeds represented, with the most frequent being Miniature Dachshunds (MDs) (n=35, 70%). Twenty-three dogs were on immunosuppressive therapy for less than 105 days, because of death during immunosuppressive therapy (n=8), loss to follow-up (n=8), or because they underwent splenectomy as an alternative therapy (n=7) (Fig. 1).

Fig. 1.

Sankey diagrams illustrating the study design of 50 dogs diagnosed with precursor-targeted immune-mediated anemia. Outcomes are written horizontally in each node with the height representing the number of dogs.

Characteristics of responders and non-responders

Twenty-seven dogs that could be followed for more than 105 days were evaluated for the effectiveness of the treatment. The median age was 10 years (range: 3–14 years) and there were 16 females (8 intact and 8 neutered) and 11 males (all neutered). The breeds were 17 MDs, 3 Toy Poodles, 2 mixed-breed dogs, and 1 each of Beagle, Labrador Retriever, Papillon, Shih Tzu, and Wire Fox Terrier. Seven dogs (26%) had been started on immunosuppressive therapy by the referring veterinarians; this comprised corticosteroids (n=7), azathioprine (n=3), cyclosporine (n=3), mycophenolate mofetil (n=4), and recombinant human erythropoietin (n=1). The median duration of immunosuppressive treatment at the referring hospitals was 21 days (range: 3–124 days) and 8 dogs received a median of 1.5 blood transfusions (range of the number of transfusions: 1–14).

On the first visit to our hospital, all dogs showed good consciousness levels and appeared tolerant of anemia. Thoracic radiography and abdominal ultrasonography showed the following findings: hepatomegaly (n=11), splenomegaly (n=8), gallstones (n=4), small amount of ascites (n=4), intra-gastric foreign bodies (n=2), mild microhepatia (n=2), and mild prostate calcification (n=1). No cases revealed any thrombus by ultrasonography.

After the first visit, all dogs were treated with prednisolone at a median dose of 2 mg/kg/day (range: 1.3–3 mg/kg/day). Twenty-two dogs also received immunosuppressive agents as follows: cyclosporine (n=14; median dose: 7.5 mg/kg/day, range: 5–10.5 mg/kg/day), mycophenolate mofetil (n=4; median dose: 27 mg/kg/day, range: 24–30 mg/kg/day), and azathioprine (n=2; median dose: 2.4 mg/kg/day, range: 2–2.7 mg/kg/day). Over 105 days of treatment, 18 dogs were classified as immunosuppressive responders and 9 as immunosuppressive non-responders. The median duration of immunosuppressive therapy was 224 days (range: 112–2,382 days) for all cases. The median time to observe a therapeutic response was 29 days in the immunosuppressive responders (range: 6–126 days), and most dogs responded to treatment within 60 days (16 of 18 dogs), but 2 required 93 and 126 days, respectively (Fig. 2). Three of 18 immunosuppressive responders relapsed at 82, 212, and 233 days after the response date, respectively. All dogs relapsed when immunosuppressive therapy was being tapered. Three out of the non-responders continued immunosuppressive therapy until the last visit to our hospital, and 6 had their medication discontinued because of side effects or at the owner’s request. The median duration of immunosuppressive therapy in the non-responders was 154 days (range: 112–268 days).

Fig. 2.

Stacked bar graph showing the cumulative percentage of responders to long-term immunosuppressive therapy per month in 28 dogs with precursor-targeted immune-mediated anemia. White=immunosuppressive non-responders; black=immunosuppressive responders.

Clinicopathological data of responders and non-responders

In 22 of the 27 dogs, CBC was evaluated within 24 hr before bone marrow examination. In the remaining 5 dogs, CBC was evaluated 3–4 days prior to the bone marrow examination. All dogs were severely anemic (median PCV 18%, range: 8–27), with low reticulocyte counts (median reticulocytes 21.9 × 10³/μL, range: 0–50.5). Autoagglutination test results were negative in all 9 dogs tested, and 11 of 27 dogs that underwent Coombs testing were negative. Spherocytes were observed in 4 dogs, but none of these fulfilled the criteria of ≥5 spherocytes/×100 oil immersion field [5]. No dogs showed evident pancytopenia but 4 had bicytopenia: 1 showed mild neutropenia and 3 dogs had mild thrombocytopenia with anemia. Seven dogs had mild-to-moderate leukocytosis (range: 17.1–31.4 × 10³/µL) without a left shift and 8 dogs showed thrombocytosis (range: 494–889 × 10³/µL).

BM aspiration was performed in all dogs and core biopsies were obtained from 20 dogs. Aspiration was difficult in 1 dog (dry tap), which showed severe BM fibrosis on core biopsy. Fourteen dogs had erythroid hyperplasia, 8 had hypoplasia, and 4 had normoplasia. Regarding GMR and EMR, 15 dogs were within the RI and 5 were above the RI for GMR, whereas 7 dogs were within the RI and 19 were above the RI for EMR. The median macrophage ratio was 0.2% of all nucleated cells, but the ratio in 1 dog was 4.2%, which was above the RI [6]. Rubriphagocytosis was detected in 8 dogs. In accordance with a previous report [18], BM polychromatophils were present in 19 dogs, despite the peripheral blood samples showing non-regenerative anemia. Among the 20 core biopsies evaluated, 13 dogs had fibrosis that was graded as mild, moderate, or severe in 6, 2, and 5 dogs, respectively.

Predictors of responses to immunosuppressive therapy

The characteristics and clinicopathological data of the immunosuppressive responders and non-responders are shown in Table 1. Based on the characteristics at admission, immunosuppressive non-responders were prone to be older than responders (P=0.049). However, there were no significant differences in relation to sex, neutering status, body weight, and CBC results. EMR and GMR were significantly related to the response in BM, which was significantly lower in immunosuppressive responders compared with non-responders (P=0.026 and 0.046). A logistic regression analysis was conducted to determine whether age, EMR, and GMR were predictors of treatment response. Our results found that EMR and GMR were significantly associated with treatment response (age P=0.100 vs. EMR P=0.015 vs. P=0.024). The AUC of EMR was higher than that of GMR (EMR AUC=0.74 vs. GMR AUC=0.71). Therefore, we focused on EMR and a cutoff value of 0.17 was obtained based on the highest Youden Index (Fig. 3). These cutoff values were associated with a sensitivity and specificity of 70.6% and 77.8%, respectively, to predict a treatment response.

Table 1. Characteristics and clinicopathological data for immunosuppressive responders and non-responders.

Variable	Reference interval	Respondersa	n	Non-respondersb	n	P value
Sex	-	Male 8 / female 10	18	Male 3 / female 6	9	0.692
Neutering status (intact/neutered)	-	Intact 6 / neutered12	18	Intact 2 / neutered 7	9	0.676
Age (years)*	-	9 (3–14)	18	12 (5–14)	9	0.049
Body weight (kg)	-	6.41 (4.1–30.5)	18	5.4 (3.48–7.66)	9	0.198
PCV (%)	37–55	16.5 (8–27)	18	18 (12–25)	9	0.435
Reticulocytes (×103/μL)	10–110	19.9 (0–50.5)	18	22.2 (5.5–38.9)	9	0.338
Platelets (×103/μL)	148–484	366 (49–889)	18	307 (136–843)	9	0.41
CR%	0–1.0	0.26 (0–0.72)	18	0.34 (0.06–0.85)	9	0.202
Result of bone marrow aspiration
Polychromatophils (+ / −)	-	Observed 12 / not observrd 5	17	Observed 7 / not observrd 2	9	1
Rubriphagocytosis (+ / −)	-	Observed 5 / not observrd 12	17	Observed 3 / not observrd 6	9	1
EMR*	0.02–0.12	0.14 (0.06–0.50)	17	0.42 (0.06–0.80)	9	0.026
GMR*	0.04–0.15	0.05 (0.01–0.76)	17	0.05 (0–0.11)	9	0.046
Macrophages (%)	<1	0.2 (0–0.8)	17	0.2 (0–4.2)	9	0.235

CR%, corrected reticulocyte %; ME ratio, myeloid-to-erythroid ratio; EMR, erythroid-maturation ratio; GMR, granulocyte-maturation ratio. Data expressed as median (range) or number. *Significant difference between immunosuppressive responders and non-responders (P<0.05). ^aResponders defined as absolute reticulocyte concentration increased to ≥60,000/μL; ^bNon-responders defined as absence of increased absolute reticulocyte concentration (<60,000/μL) up to 105 days after starting treatment.

Fig. 3.

Receiver operating characteristic curves for the cutoff point of the erythroid-maturation ratio.

Complications of immunosuppressive therapy

Of the 50 PIMA-affected dogs included in this study, 8 died while on immunosuppressive therapy, and 2 died because of progressive anemia 14 and 33 days after starting immunosuppressive therapy. In these two cases, the medication was not tapered. The remaining 6 cases died of complications including pancreatitis (n=3), pneumonia (n=2), and neurological disorder (n=1). One of the 2 dogs that developed pneumonia also had megaesophagus. One dog that developed a neurological disorder presented with nystagmus and seizures and died 6 days later with a loss of consciousness. At the onset of complications, all dogs were receiving prednisolone. The details of treatments and outcomes are shown in Table 2.

Table 2. Complications of immunosuppressive therapy with precursor-targeted immune-mediated anemia.

Complication	Duration of immunosuppressive therapy at the time complication occurred (days)	Dosage at the time of complication (mg/kg/day)	Response for immunosuppressive therapy	Outcome
Aspiration pneumonia	7	Pre (2) and MMF (20)	NE	Died on the same day
Pancreatitis	15	Pre (2) and CyA (8.7)	NE	Died 13 days later
Pancreatitis	22	Pre (2.3) and CyA (11.5)	NE	Died 4 days later
Neurological disorder	29	Pre (1.9) and CyA (20)	NE	Died 6 days later
Pancreatitis	74	Pre (1.6) and MMF (14.8)	NE	Died the next day
Orbital abscess	77	Pre (1.4) and CyA (10)	NE	Survived
Pneumonia	102	Pre (1.4) and Lef (4.3)	NE	Died the next day
Hepatic abscess	114	Pre (1.4) and CyA (9.4)	Non-responder	Died the next day
Aspiration pneumonia	116	Pre (2) and MMF (26.8)	Responder	Died 10 days later
Pyothorax	155	Pre (0.2) and Lef (3.2)	Non-responder	Died the next day
Prostate abscess	167	Pre (1.2) and MMF (27)	Responder	Survived
Pancreatitis	318	Pre (2) and MMF (30.4)	Responder	Died 30 days later
Pyometra	1,491	MMF (12.5)	Responder	Survived

CyA, cyclosporine; MMF, mycophenolate mofetil; Lef, leflunomide; Pre, prednisolone; NE, not evaluated.

Of the 27 dogs with continuous immunosuppressive treatment for more than 105 days, complications occurred in 6 dogs (22%). Of these, 4 were responders and 2 were non-responders. At the onset of the complications, 5 out of 6 dogs were treated with prednisolone and immunosuppressive drugs and 1 dog was treated with mycophenolate mofetil alone. Immunosuppressive therapy was discontinued on the day of complication onset in 4 dogs (hepatic abscess, pyothorax, prostate abscess, and pyometra). One dog developed aspiration pneumonia and died during the tapering of prednisolone. One dog developed pancreatitis 49 days after resuming immunosuppressive therapy following a relapse of PIMA. Therefore, this dog continued immunosuppressive therapy after the onset of pancreatitis. Regarding the outcomes, 2 dogs recovered from complications but the other 4 died within 30 days of diagnosis. The dog with prostate abscesses discontinued immunosuppressive therapy and received blood transfusions as necessary. The dog with pyometra resumed mycophenolate mofetil 18 days after hysterectomy.

DISCUSSION

Thromboembolic events and CR% were not shown to indicate treatment response or to be prognostic factors in this retrospective study. Assenmacher et al. and Woolhead et al. previously showed that thromboembolic events and CR%, as a measure of the degree of hematopoiesis, were prognostic factors for survival in dogs with non-regenerative anemia that received immunosuppressive therapy as first-line treatment [1, 18]. Therefore, we investigated these variables in our study population. In the present study, only 1 dog had portal vein thrombus and died. However, the cause of thrombus was suspected to be associated with pancreatitis. We also compared the median CR% between immunosuppressive responders and non-responders. Woolhead et al. reported that a CR% >0.2 was associated with decreased mortality at 3 months and 12 months in NRIMA dogs, and therefore, we expected a lower CR% in the non-responders [18]. However, there were no differences in median CR% between responders (0.26) and non-responders (0.34). The median CR% reported by Woolhead et al. was 0.7 compared with 0.30 in the current study. This apparent discrepancy might be related to differences in the case populations between the studies. There were also differences between the studies in terms of tests used to examine immune-mediated erythrocyte destruction in the peripheral blood. Woolhead et al. reported positivity rates for the saline agglutination test, Coombs test, and spherocytosis of 54%, 8%, and 19%, respectively; however, these were all negative in our study [18]. These differences indicate that the cases in the study by Woolhead et al. had more clinical findings indicating an immune response to peripheral erythrocytes compared with our cases, which may have contributed to the differences in CR% between the two studies.

In contrast, the present results suggested that EMR might be a predictor of treatment response. We observed a median EMR of 0.14 in the immunosuppressive responders and 0.42 in the non-responders. A higher EMR indicates a larger number of immature cells, which suggests that immature erythroblasts were more abundant in immunosuppressive non-responders, indicating more pronounced maturation arrest. To the best of our knowledge, no previous reports have examined EMR and treatment response in relation to non-regenerative anemia. Lucidi et al. reported the presence of immunoglobulin G (IgG) and phosphatidylserine (PS) in erythroid lineage cells at each maturation stage within bone marrow of dogs with PIMA. Their study demonstrated that IgG and PS positive cells were more likely to be present in the mature rubricyte cells [7]. This result suggested that mature nucleated red blood cells are susceptible to destruction by the immune system. Thus, the high EMR can be the result of active regeneration of the erythroid lineage to compensate for immune-mediated destruction of mature nucleated red blood cells. Although this study did not include cases with marked dysplastic features of bone marrow cells, high EMR is known to be a feature of primary myelodysplastic syndrome with refractory anemia [17]. Tani et al. reported that they suspected the possibility of myelodysplastic syndrome in MDs that were resistant to immunosuppressive therapy because of higher platelet counts and higher frequencies of hyper-segmented neutrophil and spindle shape platelet in peripheral blood compared to responders [14]. If non-regenerative anemia in MDs is difficult to distinguish between PIMA and myelodysplastic syndrome on bone marrow cytology, the higher EMR in the non-responders cold be due to the inclusion of myelodysplastic syndrome cases. Further research is warranted to determine whether EMR can be a prognostic factor, including approaches to pathophysiology.

Overall, the treatment response in the current study was 29 days, which was similar to previous reports by Assenmacher et al. and Woolhead et al., reporting 29 and 31 days, respectively [1, 18]. The distribution of treatment response time showed that 9 of the 18 dogs responded within 30 days, 7 within 30–60 days, and the remaining 2 responded by 93 and 126 days, respectively. Therefore, to distinguish between responders and non-responders, treatment should be continued for at least 60 days, even if no response is detected by 30 days.

In this study, complications were observed over the entire study period from initial treatment to long-term immunosuppression. Pancreatitis (n=4), abscesses (n=3), and pneumonia (n=3) were observed in multiple cases. All three pneumonia cases in this study were diagnosed in MDs, with aspiration pneumonia secondary to megaesophagus in two of the three cases. In a retrospective study in Japan, megaesophagus was significantly overrepresented in MDs [10]. Higher numbers of aspiration pneumonia cases associated with megaesophagus might have been observed in this study compared with previous studies because of the high number of MDs. Likewise, pancreatitis was also reported in immune-mediated diseases treated with protocols similar to PIMA [4]. While the relationship between pancreatitis and immune suppressive therapy or non-regenerative anemia is unknown, all dogs diagnosed with pancreatitis suffered from persistent anemia and were prescribed prednisolone at a dose of more than 2 mg/kg. Therefore, pancreatitis may be a complication requiring attention, especially during periods of uncontrolled anemia.

Additionally, pancreatitis and pneumonia were reported as complications in previous PIMA/NRIMA retrospective studies [1, 18], our study is the first to report the presence of infection, including abscesses. These complications were observed in cases receiving immunosuppressive treatment from 77 to 167 days. In a report of opportunistic fungal infections of the skin, the median diagnosis of opportunistic infection was reported at a median of 43 days after the immunosuppressive treatment was initiated [9]. Although at different sites of infection, the abscesses and pyothorax observed in this study were most likely opportunistic infections related to the underlying immunodeficiency caused by prolonged immunosuppressive therapy. Furthermore, 2 of the 4 dogs that developed a bacterial infection died because of complications, suggesting that opportunistic infections may have a significant impact on prognosis. Therefore, it is important to consider the following two points when planning a treatment course for PIMA: first, most dogs that responded to immunosuppressive treatment had an effective response within 60 days; and second, we should carefully consider treatment lasting more than 60 days, especially with regard to the onset of opportunistic infections associated with immunodeficiency.

This study had several limitations. First, the study population was small and had a breed bias. Although there were 65 suspected cases of PIMA, 15 dogs were excluded because BM examination was not performed. Thus, some comparisons may have lacked the required statistical power to show a significant difference. However, most previous reports focused on immunosuppressive responders, whereas the current study is the first to observe both responders and non-responders to immunosuppressive therapy for PIMA over a span of a certain period. Second, the study was limited by its retrospective nature. Because 26% of the dogs were referred to our hospital after the initiation of treatment, we were unable to standardize the time interval between the onset of treatment and bone marrow testing, which could have affected the findings.

In conclusion, a novel finding of our study is that an EMR <0.17 is a positive prognostic factor. EMR can be calculated from the initial bone marrow examination, and it may be useful for predicting treatment responses and initial treatment planning. Furthermore, few dogs improved from non-responders to responders after 60 days of treatment whereas more complications, including suspected opportunistic infection, were observed after 70 days. Therefore, we concluded that treatment plans should be reconsidered if immunosuppressive therapy is ineffective, despite sustained administration for 60 days. The effectiveness of alternative treatments such as splenectomy [2] should also be examined to reduce the risk of complications associated with the long-term use of immunosuppressive therapy.

CONFLICTS OF INTEREST

The authors declare no conflicts of interest associated with this manuscript.