Introduction

Immune-mediated hemolytic anemia (IMHA) in dogs is a common autoimmune disease with a high mortality rate (24–58%) and financial investment for owners.^1–5 IMHA is defined as an immune-mediated (antibody and/or complement driven) destruction of red blood cells (RBCs).^1–4,6,7 Relapses are known to occur, but risk factors for relapse are poorly defined.^8,9 A systematic review of IMHA determined a relapse rate of 6–15% in dogs.^4,10 A median relapse time of 112 days was noted in 18 relapsing dogs.¹¹ Another study found a 13% relapse rate with most relapses occurring within 2 to 21 days after discharge, although one patient relapsed 1300 days after initial diagnosis.¹

The lack of a standardized definition of IMHA relapse has complicated the understanding of this subject. One study defined a relapse as a packed cell volume (PCV) of <25% at any time.¹ The median time to relapse in this study was 7 days, which may have indicated failure to respond to initial therapy rather than a true relapse.¹ A subsequent study avoided this confounding factor by defining a relapse as a drop in PCV after improvement or a complete recovery followed by a positive Coombs test or erythrocyte fragility test.¹¹ Other research mentions relapse rates in their study population but do not define what they consider a relapse.^12,13

Given the inconsistent definitions and nonspecific measurements of IMHA relapses in past studies, further research and more precise evaluations of IMHA relapses are needed. Beyond basic descriptive analysis of relapses, studies looking into relapses and their various risk factors are scant. By understanding the risk factors for relapse, clinicians could identify patients at higher risk and more closely monitor them. The authors hypothesized that patients with more severe anemia on presentation (lower PCV at diagnosis, more transfusions administered) would be at greater risk of relapse. The authors also hypothesized that patients with increased hemolysis (demonstrated by increased total bilirubin) would be at greater risk of relapse. By identifying factors that predict an IMHA relapse, better monitoring regimens may be instituted for patients at a higher risk. Identification of factors associated with IMHA relapse may also direct future therapeutic interventions if they are also determined to be causative.

Materials and Methods

Records between January 1, 2005, and December 31, 2019, from one specialty referral hospital were reviewed for any instance of the terms “immune-mediated hemolytic anemia,” “IMHA,” “autoimmune hemolytic anemia,” and “AIHA.” Feline patients were excluded. Primary care records were unavailable for nearly all patients and thus not reviewed. In accordance to the American College of Veterinary Internal Medicine definition of IMHA, patients were considered to have IMHA if their PCV at diagnosis was less than 30% and if two or more of the following were identified: spherocytes, positive Coombs test, hemoglobinuria, hyperbilirubinemia, or positive slide agglutination test (micro or macro).¹² The presence of spherocytes was confirmed by a laboratory technician, veterinarian, or board-certified pathologist. Slide agglutination was confirmed by either a laboratory technician or veterinarian. Patients were excluded if another cause of anemia was found (trauma, anemia of chronic disease, primary liver disease, or chronic kidney disease). Cases of associative IMHA (i.e., Babesia, neoplasia, drug reactions) and precursor-IMHA (also known as precursor-targeted immune-mediated anemia or pure red cell aplasia) were excluded as well. Patients were considered to have relapsed if these conditions were met: a previously documented PCV >30%, greater than 1 mo after initial diagnosis, and a decrease in PCV greater than 10% (i.e., 30% to <20%). This decrease excluded incidents of trauma, surgery, or neoplasia. A PCV of greater than 30% was used to define adequate control and management of the disease. If the patient’s PCV never exceeded 30% after discharge, they were considered to have inadequately controlled IMHA and thus could not relapse from an active disease. One month was chosen as the majority of clinicians did not typically decrease medications during the initial month. The records were reviewed for number of transfusions administered to manage IMHA, and time to euthanasia, death, or loss to follow-up. Because of the retrospective nature of this study, monitoring schedules and tapering regimens differed between clinicians.

Results

Of the records reviewed, 163 dogs met the criteria of IMHA diagnosis. There were 13 dogs who relapsed, indicating a relapse proportion of 7.9%. Seventy-six dogs were censored because they were euthanized or died before relapse, and 74 dogs were censored as a result of being lost to follow-up. The median (min–max) follow-up time was 534 (1–3507) days. The overall probabilities of relapse (95% confidence interval [CI]) at 3 and 12 mo were 0.05 (0.02–0.13) and 0.11 (0.06–0.22), respectively (Figure 1).

View Figure

• Download as Powerpoint
• Download Figure

The probability of relapsing by 1 yr (95% CI) for dogs that did not need one or more transfusions was 0.05 (0.02–0.25), for dogs that needed one or more transfusions was 0.14 (0.07–0.27), and for dogs who received two or more transfusions was 0.20 (0.09–0.42). However, there was not a significant association between risk of relapse and having any (P = .101) or two or more transfusions (P = .191). Sex was not significantly associated with risk of relapse (P = .962). The probability of relapsing by 1 yr (95% CI) for male neutered dogs was 0.19 (0.07–0.43), and for female spayed dogs was 0.10 (0.04–0.24). These findings are displayed in Table 1.

TABLE 1 Three-Month and Twelve-Month Relapse Probabilities by Categorical Factors

View Fullscreen

Hazard ratios (95% CI) for relapse risk for PCV, total bilirubin, and age at diagnosis were 0.95 (0.86–1.04), 1.17 (1.06–1.28), and 1.13 (0.95–1.34), respectively. Only an increase in total bilirubin was significantly associated with an increased relapse risk. There was a 17% increase in the probability of relapse per 1 mg/dL increase in total bilirubin. Table 2 provides further details about these continuous factors.

TABLE 2 HRs for Continuous Factors

View Fullscreen

Discussion

The primary aim of this study was to identify risk factors for dogs developing a relapse of IMHA. This study’s overall relapse rate of 7.9% falls within the range of previously reported relapse rates of 6–15%.^1,4,10,11,14 Increases in total bilirubin were significantly associated with increased relapse risk. Patients with a higher total bilirubin at diagnosis may represent a population with a more aggressive process, inadequate response to immunosuppressants, faster hemolysis, or delayed medical intervention.

Initially, we suspected that patients requiring more transfusions could indicate a slower response to medications, faster hemolysis, or lower PCV at diagnosis. Transfusions were determined by the number of units packed RBCs purchased by the owner if transfusion logs were absent or incomplete. Larger patients may have had a greater transfusion requirement owing to a larger volume requirement. Conversely, smaller patients may have been able to receive multiple transfusion volumes from 1 unit. Cross-matching was done at the discretion of the clinician, and results were inconsistently recorded. Thus, it was difficult to investigate the influence of transfusion compatibility on the number of transfusions administered. Whether patients were transfusion naive or previously transfused at diagnosis may be an area for future study. It is likely that all of these factors played a role in not finding significance for increased transfusion needs with relapses. However, patient weight and size were not considered in the data analysis. The majority of transfusions were done with commercially available packed RBC units. Of those dogs who required two or more transfusions, it was estimated 20% of them would relapse within the first year. This is a substantial increase from the estimated 5% one-year relapse rate of patients who did not require any transfusions. The same can be said when looking at patients who needed less than two transfusions (7% one-year relapse rate).

“Typical” treatment regimens for IMHA use prednisone/prednisolone and an additional immunosuppressant agent (such as microemulsified cyclosporine, mycophenolate mofetil, azathioprine, and leflunomide).^1,2,7,10,11 Owner compliance on administering medications, follow-up, and agreement to further diagnostics may have played a role in relapse rates. The undetermined extent of owner adherence to prescribed medication protocols was another study limitation. Determination of compliance is difficult in retrospective studies and was not considered as an independent factor.

Previous studies have shown that females are at a higher risk of developing IMHA and immune-mediated diseases in general.^1,3,12 However, our data did not show a significant sex predilection for relapsing, and the estimated probability of relapsing by 3 or 12 mo was lower in spayed females than neutered males. Spaying could theoretically remove some hormonal influence on the prevalence of immune-mediated diseases. Older patients are more likely to have concurrent diseases that can affect IMHA control and thus relapse. Age at diagnosis, however, was not found to be a significant risk factor in relapsing.

Cocker spaniels and dachshunds were the most commonly affected breeds in our data set, which aligns with previous studies.^1,3,5,11 Chihuahuas were also a commonly affected breed in our data, which is a new finding not previously reported. The number of relapses for each breed was not large enough to determine a breed predilection for relapsing.

The retrospective nature of this study and the large number of patients lost to follow-up also limit this study’s findings. It is possible that some of the patients that were lost to follow-up went on to relapse. Given the frequent need for rechecks and cost of hospitalization during the initial IMHA crisis, financial constraints may have been a major factor of owners choosing long-term follow-up with their primary care veterinarians rather than choosing to follow up at a specialty clinic. The wide time frame (14 yr) of our study covered a period of many changes in treatment modalities as new research became available and clinician preference changed.

The main limitation of this study was the sample size of 13 relapses, which could have caused type II errors. In the future, a similar study with a larger relapse population may provide further clarity as to whether transfusions are associated with increased risk of relapse. Further data and other clinical studies would be needed to determine whether there is a breed disposition to relapses or whether other diagnostic factors may be a relapse risk factor.

This study adds further knowledge about IMHA, specifically in cases who relapse. Mortality rates for IMHA are high (24–70%), with the majority of previously published data looking at survival to discharge, 1 mo survival, and 1 yr survival.^{1,3,4,6,11,14} Unfortunately, survival to discharge is often only the first step in a much longer management process. Owners need to remain vigilant to clinical signs that signal a sudden drop in RBC status.¹⁵ Our study shows that patients’ total bilirubin levels at diagnosis are significantly associated with their likelihood to develop relapses. It is easy to describe these findings as the more “severe” a presentation of IMHA, the more likely a patient will relapse. However, this line of thought is debatable, as other unidentified factors may be responsible. Further study is needed to determine what these early signs could be.

IMHA can be a financially and medically intensive disease that requires frequent monitoring and medication adjustments. By identifying patients who may be at higher risk of relapsing, better client education, long-term prognosis, and monitoring can be done. Previous studies of IMHA looked at survival to discharge, evaluated a limited time after discharge,^11,14 or undertook a long-term analysis of various medical therapies.^1,8,10 Patients at higher risk of relapse may require closer monitoring and stricter control of potential triggers.

Conclusion

This study delved further into specifics of IMHA relapse. Patients with increase total bilirubin at diagnosis will have an increased risk of relapse. Larger study populations in the future would provide narrower CIs for a number of factors and thus more definitive conclusions.

Given the retrospective nature of this study, Institutional Animal Care and Use Committee approval was not needed. All patients described in this study were clinically managed according to contemporary standards of care.