Dilated cardiomyopathy in cats: survey of veterinary cardiologists and retrospective evaluation of a possible association with diet

Abstract

Introduction/objectives:

The objectives were to conduct a survey of cardiologists on their recent experiences with cats that have dilated cardiomyopathy (DCM) and to retrospectively review individual cases of feline DCM.

Animals, materials and methods:

Part one: A survey was distributed to cardiologists with questions regarding caseload and clinical management of cats with DCM diagnosed over the past two years. Part two: Cardiologists completing the survey were invited to submit data from cats recently diagnosed with DCM. Data on signalment, clinical signs, diet, echocardiographic measurements and outcome were recorded.

Results:

Part one: From 52 completed surveys, many cardiologists responded that measuring and supplementing taurine and recommending a diet change in cats with DCM are common practices. Few (15%) cardiologists reported an increase in the number of feline DCM cases over the past two years, although some had cases that improved even if taurine deficiency was not present. Part two: Twenty of 37 (54%) cats ate low pea/lentil (low PL) diets, and 14/37 (38%) ate high PL diets at the time of diagnosis; three had incomplete diet information. Two of 13 cats (15%) in which taurine was measured had levels below the reference range. After adjusting for other variables, cats eating high PL diets that changed diets after diagnosis had a significantly longer survival time than that of cats eating high PL diets that did not change diets after diagnosis (P = 0.025).

Conclusions:

Additional research is warranted to determine whether there could be a possible association between diet and DCM in cats.

Introduction/objectives

Since 2014, the United States Food and Drug Administration (FDA) has received reports of dilated cardiomyopathy (DCM) in pets with a potential link to certain diets [1]. Many of the cases had been fed diets which were grain-free or contained pulses (e.g. peas, chickpeas and lentils), and to a lesser extent, potatoes [1]. As of September 2020, more than 1100 dogs with DCM had been reported to the FDA [2]. Four recent studies focused on apparent diet-associated DCM: one in Golden Retrievers with taurine deficiency and three in dogs of various breeds (most of which were not taurine-deficient) [3–6]. Although the cause of this possible diet-associated DCM remains unknown, these studies noted an association with diets that were grain-free or contained peas, lentils, beans or chickpeas as primary ingredients and found that cardiac dysfunction often improved after a diet change and, in many cases, taurine supplementation [3–6]. This possible association between diet and DCM has been mainly identified in dogs, but the FDA received reports of 14 cats with possible diet-associated DCM between 2014–April 2019 as well [1].

Historically, DCM was a common disease in cats, but in 1987, an association between taurine deficiency and feline DCM was identified, leading to an increase in taurine levels in commercial cat foods and a marked reduction in cases of feline DCM [7,8]. Anecdotally, most cats diagnosed with DCM since 1987 have had normal taurine concentrations, although taurine deficiency-induced DCM can still occur in cats eating unconventional diets (e.g. vegetarian/vegan or homemade).

Feline DCM, when it does occur, is a deadly disease. In a retrospective study of cats with DCM diagnosed between 2001 and 2010, the median survival times were only 12 days in cats that did not receive pimobendan and only 49 days even in cats that did receive pimobendan [9]. Given the dramatic reduction in prevalence of taurine deficiency and feline DCM since the 1980s, current practices of evaluating and managing patients with this disease are unknown, including how frequently taurine analysis is performed, how many cats are taurine deficient, how many cats change diets after diagnosis and if clinical or echocardio-graphic improvement occurs after medical treatment and diet change, taurine supplementation or other treatment. Although the number of cases of feline DCM in recent years has not been identified, some veterinary cardiologists have been anecdotally reporting a subjective recent increase in the number of cats being diagnosed with DCM.

Subjectively, many cats in the United States and Canada appear to be eating grain-free or high pea/lentil (high PL) diets similar to those associated with DCM in dogs; these diets have grown immensely popular in recent years. For example, in 2018, grain-free diets were reported to represent 46% of the commercial pet food market [10]. Given the fact that cats have more unique nutritional requirements than dogs (e.g. taurine is an essential nutrient, absolute requirement for niacin and low delta-6 desaturase activity), cats could be more or less susceptible than dogs to possible dietary factors associated with DCM, depending on the aetiology (e.g. nutritional deficiency or toxicity) [11]. If some cats have DCM that is diet-associated but not caused by taurine deficiency, it is not yet known whether they can improve with diet change, as seen in some dogs with diet-associated DCM [3–6].

To help fill in these information gaps on current case numbers as well as practices for evaluation and treatment, diet and outcome of cats with DCM, the objectives of this study were to 1) conduct a survey of cardiologists on their experiences with cats with DCM in the last two years and 2) to retrospectively review individual cases of feline DCM, focussing on diet, echocardiographic measurements, taurine status and outcome after diet change or taurine supplementation. We hypothesised that there would be a perceived increase in frequency of DCM diagnoses in cats over the past two years, many of these cats would be eating high PL diets, most would have normal taurine concentrations, and cats eating high PL diets would improve with diet change.

Animals, materials and methods

Part one: cardiologist survey

An electronic surveyⁱ concerning feline DCM was developed and distributed to the American College of Veterinary Internal Medicine Cardiology Diplomates via its listserv (which had 362 members at the time of writing this article). The questions in this survey focused on how cardiologists generally manage cats with DCM and their practices’ caseload of feline DCM over the past two years (Table 1). Cardiologists were also asked if they would be willing to provide cases for the retrospective portion of the study (part two). The survey was available for completion between 16th April 2020 and 15th June 2020, with two reminders sent after the initial request. Surveys were considered complete if all questions available to a respondent were answered and partially complete if at least 66% of questions available to them were answered. Both complete and partially complete surveys were included in the analysis. Summary statistics were analysed, with results presented as total numbers and percentages.

Table 1.

Summary of responses from 52 board-certified veterinary cardiologists to a survey on feline dilated cardiomyopathy (DCM).

Survey question	Total responses
Regarding general approach to clinically managing feline DCM
Measure plasma or whole blood taurine concentrations in cats with DCM
Always	15
Sometimes	15
No	15
Other	6
Did not answer	1
Supplement taurine in cats with DCM
Yes — if proven to be taurine deficient	16
Yes — even if not proven to be taurine deficient	25
No	4
Other	6
Did not answer	1
Change the diet of cats when diagnosed with DCM
Always	5
Sometimes	28
No	9
Other	9
Did not answer	1
Regarding personal caseload of feline DCM seen over the last two years
Approximate number of cases of feline DCM diagnosed per year
0 cases	4
1–2 cases	26
3–5 cases	17
6–10 cases	5
>10 cases	0
Perception of the number of cases of feline DCM in the last 2 years versus previous years
Numbers have increased in the last two years	8
Numbers have not changed in the last 2 years	40
Numbers have decreased in the last two years	0
I do not know/unable to answer	4
Have seen cases of feline DCM suspected to have a nutritional aetiology (taurine deficiency or other nutritional causes)
Yes	20
No	32
Regarding personal caseload of suspected diet-associated feline DCM seen over the last two yearsa
Approximate number of cases of suspected diet-associated feline DCM seen in the last two years
1–3 cases	14
4–6 cases	4
7–10 cases	2
Cases of suspected diet-associated feline DCM that had low taurine concentrations (whole blood or plasma)
Yes – all cases	1
Yes – some cases	2
No – none were deficient	8
Taurine was not measured	9
Cases of suspected diet-associated feline DCM that showed significant clinical or echocardiographic improvementb
Yes – if cats were taurine-deficient and taurine was supplemented	3
Yes – if cats were eating a non-traditional diet and diet was changed	4
No	7
Other	10

^aThese questions were only answered by cardiologists who positively responded to having seen cases of feline DCM suspected to have a nutritional aetiology, resulting in a total of 20 responses.

^bThis question allowed respondents to select more than one answer for their response.

Part two: a retrospective study

Cardiologists who agreed in the survey to contribute cases were contacted and asked to submit records on individual cats diagnosed with DCM during the last two years for review. Cases from two of Tufts University’s small animal hospitals were also included.

Medical records, including echocardiographic reports, were reviewed for potential inclusion in the retrospective study. Echocardiographic criteria used to define DCM in this study were based on previous studies and textbook chapters on feline DCM [7–9,12–19]. To be considered for inclusion in this study, cats had to have a diagnosis of DCM by the attending cardiologist and meet the study’s echocardiographic criteria for DCM, which included left ventricular internal dimension at end-systole ≥1.2 cm and fractional shortening ≤28%, with subjective global left ventricular systolic dysfunction. The diagnosis of congestive heart failure (CHF) was based on the presence of pleural effusion or pulmonary oedema on thoracic radiography or thoracic ultrasound, in combination with a diagnosis of DCM based on echocardiography. For cats with CHF or arterial thromboembolism, the ratio of the left atrium to aortic dimensions (LA:Ao) on a short axis two-dimensional (2D) view had to be ≥ 1.6. Cats with diffuse or focal left ventricular hypertrophy were excluded (interventricular septal thickness at end-diastole and left ventricular posterior wall thickness at end-diastole both had to be ≤ 0.55 cm), as were cats with a previous diagnosis of hypertrophic cardiomyopathy. Measurements of the left ventricle were available via M-mode only from some sites, via M-mode and 2D from some sites and via 2D only from some sites. M-mode measurements were preferentially used because they were the most common measurements available. In any cat where both M-mode and 2D measurements were available and there was a discrepancy between the measurements of the two modes that resulted in a question of eligibility (n = 1), the submitting cardiologist was contacted and indicated that the 2D measurements should be preferentially used.

Records were reviewed for each eligible cat, and information was collected from the date of diagnoses on the following: signalment and vital signs; presenting clinical sign for evaluation (CHF, arterial thromboembolism, syncope, asymptomatic or other); physical examination findings; initial echocardiogram; and taurine and N-terminal pro-B-type natriuretic peptide (NT-proBNP) concentrations (quantitative or SNAP^®; if measured). Only NT-proBNP concentrations taken near the time of diagnosis were considered; all taurine concentrations were measured at a single commercial laboratory.^j Other data that were recorded included cardiac medications, measurements from the most recent echocardiogram (if more than one echocardiogram was performed), diet history and outcome.

Cats were categorised into low pea/lentil diet (low PL diet) and high PL diet groups based on their main diet at diagnosis (i.e. the diet that supplied the greatest number of calories to the cat); whether the cat changed diets or continued to eat the same diet after diagnosis of DCM was also recorded. Where possible, primary care veterinarians or cat owners were contacted for additional dietary information. Cats without follow-up information that lived less than four weeks after diagnosis and received no recommendation to change diet (n = 3) were classified as having made no diet change after diagnosis. For the purposes of the study, diets were classified as low PL when they did not contain peas, pea fractions (e.g. pea protein, pea starch and pea fibre) or lentils in the first 10 ingredients on the products’ ingredient lists; high PL diets were defined as those that contained peas, pea fractions or lentils in the first 10 ingredients on the ingredient lists. For cats in which the exact flavour of the product could not be determined, the ingredient lists for all possible products containing the known dietary information (e.g. Brand × grain-free chicken diet) were reviewed. If insufficient information was available to make categorisation possible (e.g. if the exact flavour could not be determined, and one possible flavour of a product included peas and another flavour did not include peas), the cat was excluded from diet categorisation. Ingredient lists for diets of cats seen more than six months before the time of data analysis were collected from an Internet archive website.^k

Data analysis

Data distributions were examined graphically and using Shapiro–Wilks tests before analysis. Data are presented as mean ± standard deviation for normally distributed data or median (range) for skewed data. Survival times were calculated from the time of diagnosis of DCM until the time of death/euthanasia from all causes. Cats were right-censored if they were alive at the time of analysis (20th August, 2020) or if they were lost to follow up. Survival times were compared between groups (e.g. diet groups, presence of CHF, use of individual medications and supplements or presence of arrhythmias) using Kaplan–Meier curves and log-rank tests. Variables that were significantly associated with survival in the univariable analysis (P < 0.05) were analysed further using Cox proportional hazards analysis with a forward stepwise elimination strategy with modelling stopped once all the variables left had a P-value <0.05. P-values, hazard ratios and 95% confidence intervals for the final explanatory multivariable model were calculated. All statistical tests were carried out using commercial statistical software,^l,m and P-values <0.05 were considered significant.

Results

Part one: cardiologist survey

There was a total of 56 survey responses (approximately 16% response rate based on the number of cardiologists on the listserv at the time of writing this article); four were excluded for being incomplete, resulting in a total of 52 completed surveys which were used for the final analysis. Most responses were from cardiologists practising in the United States (n = 41; 19 different states), with other responses from Canada (n = 4) and one each from Hong Kong, Italy, Japan and the United Kingdom.

Most cardiologists (30/51 or 59% of those answering this question) responded that they always or sometimes measure plasma or whole blood taurine concentrations as part of their management protocol in cats with DCM (Table 1). Other responses indicated that they measure plasma or whole blood taurine concentrations if the owner can afford the test (n = 3), if the patient is eating an ‘unusual’ diet (n = 2), if the owner does not decline (n = 1), if the patient has retinal lesions (n = 1) or if the DCM is severe (n = 1). The majority of cardiologists (41/51 or 80% of those answering this question) also responded that they supplement taurine as a form of treatment in cats with DCM whether or not taurine deficiency is documented (Table 1). Examples of other responses included supplementing taurine if it is tolerated in addition to cardiac medications (n = 1), supplementing only temporarily until taurine concentrations were available (n = 1) or supplementing if concentrations are not measured (n = 1). In terms of diet change, most cardiologists responded they always or sometimes change the diet of cats with DCM (33/51 or 65%). Others responded they only change the cat’s diet if the current diet is grain-free (n = 3), home-cooked (n = 3) or vegetarian/vegan (n = 2).

Forty-eight of 52 (92%) cardiologists reported seeing at least one cat with DCM per year in their practices, although some reported seeing up to six to 10 cats with DCM/year. Only 8/52 (15%) responded that the number of cases of feline DCM in their practices had subjectively increased over the past two years. In addition to the overall number of cats with DCM, cardiologists were also asked if they saw any cats with DCM that were suspected to be associated with diet (Table 1). Twenty of the 52 cardiologists (38%) reported seeing at least one cat suspected to have diet-associated DCM in the last two years. Of these 20 cardiologists, few (3/20 or 15%) reported documented taurine deficiency. Five of these 20 cardiologists (25%) reported seeing at least one case of DCM that was suspected to be diet-associated and showed clinical or echocardiographic improvement, although cardiologists also reported cases that did not live long enough to see improvement (n = 5), cases that had been recently diagnosed (n = 3), and cases that were lost to follow up (n = 2).

Part two: a retrospective study

Patients and diet groups

Records provided by 23 cardiology centres on 67 cats diagnosed with probable DCM were reviewed. Cats were excluded if their echocardiographic measurements did not fit within the measurements defined by the study as the parameters for inclusion or if the cat was not confirmed to have DCM (n = 24), if echo-cardiograms were missing key data required to determine eligibility (n = 4), if cats had concurrent, unregulated hyperthyroidism (n = 1) or if cats were diagnosed before 1st January 2018 (n = 1). This resulted in a total of 37 cats with DCM from 17 cardiology centres that met all the eligibility criteria and were included in the study. There were no significant differences between the five hospitals that contributed at least three cases in age, sex, weight, echocardiographic measurements, presence of CHF or arrhythmia, medications or whether diet was changed (data not shown). The primary diet for most cats was dry (n = 29), but primary diets also included canned (n = 4) and freeze-dried raw (n = 1); 20 cats were categorised in the lowPL diet group, and 14 cats were categorised in the high PL diet group based on diet at diagnosis. Three cats had an incomplete diet history and were not included in either the low or high PL diet groups but were included in summary statistics for all cats in the study.

Demographics, physical examination findings, and laboratory findings

There was a wide range of ages of cats represented (Table 2). Most cats (33/37 or 89%) had CHF at the time of diagnosis; 11/37 (30%) had a cardiac murmur auscultated, and 24/37 (65%) had a gallop. Cats in the high PL diet group had a significantly higher body weight than cats in the low PL diet group (P = 0.049), but diet groups were not significantly different in the body condition score (P = 0.509). There were no other significant differences in clinical characteristics between the low PL and high PL diet groups (Table 2).

Table 2.

Baseline comparison of signalment and physical examination findings for 37 cats with dilated cardiomyopathy diagnosed between January 2018 and August 2020. Three cats were excluded from diet categorisation owing to incomplete diet history, so the total number of cats categorised by diet is 34. Continuous data are presented as mean ± standard deviation (normally distributed data) or median (range; skewed data), and categorical data are presented as number. P-values are for comparison of the cats eating diets containing peas and lentils in the top 10 ingredients (high PL diet) compared with cats eating diets with no peas or lentils in the top 10 ingredients (low PL diet).

Variable	All cats	Low PL diet	High PL diet	P-value
n	37	20	14	–
Age (yrs)	8.8 ± 4.8	8.4 ± 4.7	8.9 ± 5.3	0.789
Sex				0.367
Male	23 (21 neutered)	11 (10 neutered)	11 (10 neutered)
Female	14 (all spayed)	9 (all spayed)	3 (all spayed)
Breed				0.325
Abyssinian	1	1	0
Domestic short/long hair	32	18	11
Ragdoll	1	0	1
Russian blue	1	0	1
Savannah	1	0	1
Siamese	1	1	0
Weight (kg)	5.1 ± 1.3	4.8 ± 1.4	5.7 ± 1.1	0.049
Body condition score (1–9)	5.6 ± 1.4	5.3 ± 1.3	5.7 ± 1.6	0.509
Presence of cachexia	10	5	4	0.845
Presenting sign				0.641
Asymptomatic	2	1	1
ATE and CHF	2	0	0
CHF	30	18	11
Syncope	1	0	1
Other	2	1	1
Heart rate (per minute)	197 ± 28	199 ± 27	189 ± 28	0.337
Respiratory rate (per minute)	48 (24–96)	42 (24–60)	50 (36–96)	0.059
Presence of murmur	11	8	3	0.255
Presence of arrhythmia
Any arrhythmia	14	6	6	0.440
Supraventricular	2	0	1	0.225
Ventricular	13	5	6	0.319
Presence of gallop	24	14	10	0.928
CHF at diagnosis	33	18	12	0.703

Key: ATE, arterial thromboembolism; CHF, congestive heart failure; PL, pea/lentil.

Thirteen cats had plasma and/or whole blood taurine concentrations measured (low PL diet group, n = 5; high PL diet group, n = 7 and insufficient diet information, n = 1). Ten cats had only whole blood taurine measured (low PL diet group, n = 2; high PL diet group, n = 7 and insufficient diet information, n = 1); two cats had only plasma taurine measured (both low PL diet group); one cat had both plasma and whole blood taurine measured (low PL diet group). Two cats had plasma taurine concentrations below the reference range (80–120 nmol/mL): one was 64 nmol/mL (low PL diet group [dry kibble]); and one was 38 nmol/mL (low PL diet group [freeze-dried raw]); whole blood taurine concentrations were not measured in these two cats with low plasma taurine concentrations. No cats had whole blood taurine concentrations below the reference range (300–600 nmol/mL).

Ten cats had NT-proBNP concentrations (quantitative or SNAP^®) evaluated near the time of diagnosis (low PL diet group, n = 6; high PL diet group, n = 3 and insufficient diet information, n = 1); all 10 cats had elevated NT-proBNP concentrations.

Initial echocardiogram

Table 3 summarises echocardiographic measurements at the time of diagnosis for all 37 cats included in the study. No significant differences were found between diet groups for any echocardiographic measurements. Only three of the 37 cats included in the study had follow-up echocardiograms (all in the high PL diet group, and all three changed diets; data not shown owing to the limited numbers).

Table 3.

Comparison of initial echocardiogram measurements at time of diagnosis for 37 cats with dilated cardiomyopathy diagnosed between January, 2018 and August, 2020. Three cats were excluded from diet categorisation due to lack of diet history. Data are presented as mean ± standard deviation. P-values are for comparison of the cats eating diets containing peas and lentils in the top 10 ingredients (high PL diet) compared to cats eating diets with no peas or lentils in the top 10 ingredients (low PL diet).

Variable	All cats	Low PL diet	High PL diet	P-value
n	37	20	14	–
LVIDs (cm)a	2.02 ± 0.44	2.08 ± 0.47	1.96 ± 0.44	0.444
LVIDd (cm)a	2.33 ± 0.45	2.37 ± 0.43	2.31 ± 0.51	0.721
IVSd (cm)a	0.39 ± 0.08	0.37 ± 0.07	0.41 ± 0.10	0.204
LVPWd (cm)a	0.39 ± 0.07	0.39 ± 0.07	0.37 ± 0.08	0.629
FS (%)a	13.77 ± 5.96	12.71 ± 5.96	15.15 ± 6.50	0.265
LA:Ao	2.20 ± 0.38	2.24 ± 0.45	2.11 ± 0.28	0.358

Key: FS, fractional shortening; IVSd, interventricular septum thickness at end-diastole; LA:Ao, ratio of the left atrial to aortic dimensions; LVIDd/s, left ventricular internal dimension at end-diastole/end-systole; LVPWd, left ventricular posterior wall thickness at end-diastole; PL, pea/lentil.

^aM-mode (n = 34), 2D (n = 3).

Medication and taurine

Over the course of their disease, cats were treated with a variety of cardiac medications, including pimobendan (n = 36), loop diuretic (n = 34), clopidogrel (n = 29), angiotensin-converting enzyme inhibitor (ACEI [n = 9]), low-molecular-weight heparin (n = 4), potassium (n = 4), spironolactone (n = 4), aspirin (n = 1), carvedilol (n = 1), sildenafil (n = 1) and sotalol (n = 1). Sixteen cats received supplemental taurine. There were no significant differences found between diet groups in terms of medication administration, but more cats received taurine supplementation in the high PL diet group (9/14 or 64%) compared with the low PL diet group (5/20 or 25%; P = 0.022).

Diet change

Ten of the 34 cats with a complete diet history had their diets changed after DCM diagnosis, whereas 24 cats did not have their diets changed. Two cats changed from a low PL diet to another low PL diet, and eight cats changed from a high PL diet to a low PL diet.

Survival

At the time of writing the article, 30 cats (81%) were no longer living: 23 were euthanised (worsening heart disease, n = 21 and arterial thromboembolism, n = 2), six died suddenly and one cat died as a result of worsening heart disease, but it is unknown whether the cat died or was euthanised; seven cats were still alive. There were no significant differences found between diet groups for the number of patients that were still alive (P = 0.335), nor for cause of death (P = 0.278). The median survival time for all 37 cats was 14 days (range 0–489 days). For the 34 cats for which baseline diet could be categorised, the median survival time for cats eating high PL diets at the time of diagnosis (n = 14) was 82 days (range 1–489 days), whereas the median survival for cats eating low PL diets (n = 20) was 10 days (range 1–273 days; P = 0.083). When cats in the high PL diet group were separated by whether or not they changed diets after diagnosis, cats eating high PL diets that changed diets after diagnosis (n = 8) had a significantly longer survival time (median 290 days, range 42–489 days) than that of cats eating high PL diets that did not change diets after diagnosis (n = 6; median 2 days, range 1–158 days; P = 0.014) or cats eating low PL diets at diagnosis (n = 20; median 10 days, range 1–273 days; P = 0.004; Figure 1). Survival time for cats that were eating high PL diets that did not change diet was not significantly different from cats eating low PL diets (P = 0.345). Using log-rank tests, use of an ACEI was associated with a longer survival time (P = 0.011), and the presence of CHF at the time of diagnosis was associated with a shorter survival time (P = 0.016). None of the other medications nor taurine supplementation were significantly associated with survival time. Cox proportional hazards analysis showed that only the diet group remained significantly associated with survival time after adjusting for the presence of CHF at diagnosis (P = 0.078) and use of ACEIs (P = 0.362), with cats eating high PL diets that changed diets living significantly longer than cats eating high PL diets that did not change diets (P = 0.025; Table 4). Cats eating high PL diets that did not change diets were not significantly different from cats eating low PL diets (P = 0.691).

Figure 1.

Kaplan–Meier survival curves for 34 cats with dilated cardiomyopathy diagnosed between January 2018 and August 2020. Cats eating a diet with peas or lentils in the top 10 ingredients (high PL diets) at the time of diagnosis that changed diets after diagnosis (n = 8; black solid line) had a significantly longer survival time than that of cats eating high PL diets that did not change diet (n = 6; grey dashed line; P = 0.014) or cats eating diets with no peas or lentils in the top 10 ingredients (low PL diets; n = 20; black dotted line; P = 0.004). Survival time for cats eating high PL diets that did not change diets was not significantly different from cats eating low PL diets (P = 0.345). PL, pea/lentil.

Table 4.

Final Cox proportional hazards model for survival time in 37 cats with dilated cardiomyopathy. Twenty cats were eating diets containing no peas or lentils in the top 10 ingredients (low PL diet), 14 cats were eating diets with peas or lentils in the top 10 ingredients (high PL diet; eight cats changed diets and six cats did not change diets after diagnosis), and three cats had incomplete diet information so could not be categorised into a diet group. Use of angiotensin-converting enzyme inhibitors was also included in the model but was not significant (P = 0.362).

Variable	P-value	Hazard ratio	95% CI for HR
			Lower	Upper
Diet group
Cats eating high PL diets that did not change diets (reference)	–	1.000	–	–
Cats eating high PL diets that changed diets	0.025	0.189	0.044	0.808
Cats eating low PL diets	0.691	0.812	0.291	2.269
Congestive heart failure (reference, no congestive heart failure)	0.078	6.204	0.814	47.265

Key: CI, confidence intervals; HR, hazard ratio; PL, pea/lentil.

Discussion

In this retrospective study, the survival curve for cats in the low PL diet group appeared to be similar to historical results for cats with DCM that were not supplemented with taurine [20] and to cats with DCM in a more recent retrospective study [9]. It is unknown exactly how many cats in this retrospective study had low taurine concentrations at the time of DCM diagnosis, because only 13/37 cats had taurine concentrations measured. However, only 2/13 cats had a taurine concentration below the reference range (both only in plasma and both in the low PL diet group). The low number of cats that had taurine concentrations measured, as well as the low number of cats that had documented low taurine concentrations, was consistent with survey results from part one of the study in which only 15 cardiologists reported always measuring taurine in cats with DCM, and only three cardiologists indicated having a feline DCM case with documented low taurine concentrations. Survival results for the current retrospective study did not show an association between taurine supplementation and survival time, suggesting a lower likelihood for a role of taurine deficiency in this group of cats. However, additional research, including prospective studies, is needed to better understand the current role of taurine deficiency in cats with DCM and, for now, measurement of plasma and whole blood taurine concentrations in cats with DCM is recommended.

Although the survival time of the high PL diet group as a whole was not significantly longer than the survival time of the low PL diet group, cats eating high PL diets that changed diets after diagnosis lived significantly longer than cats eating high PL diets that did not change diets. Cats eating high PL diets that changed diets after diagnosis also lived significantly longer than cats eating low PL diets. These findings are similar to those from a recent retrospective study of dogs with DCM in which dogs eating non-traditional diets that changed diets after diagnosis had a longer survival time than that of dogs eating non-traditional diets that did not change diets [5]. These findings for survival time support a theory that some cats in the high PL diet group had a diet-associated DCM that had a favourable outcome if diet was changed; again, it is unlikely that this association or favourable outcome is related to dietary taurine, seeing as taurine supplementation within the retrospective study group was not significantly associated with survival time. Alternatively, because the association between diet group and survival cannot show causation, this finding could indicate that cats with less severe disease lived longer after a diagnosis of DCM and therefore had a longer time during which diet could be changed. However, nearly all cats in the present study had CHF at the time of diagnosis, suggesting relatively advanced disease in most cats. Differences in survival time between diet groups could also be related to genetic factors or other environmental modifiers. The effects of diet change on survival time in the low PL diet group are unknown because only two cats in this group changed diets after diagnosis. Despite these findings of a longer survival time in cats eating high PL diets that changed diets, prospective studies which include larger numbers and follow-up echocardiography are needed to determine whether some cats with DCM could have a diet-associated form of disease similar to what appears to be occurring in dogs.

In addition to the longer survival times of cats eating high PL diets that changed diets, of cardiologists responding to the survey in part one who suspected that at least one of their cases of feline DCM had a nutritional aetiology, some reported having patients that were not taurine-deficient but showed significant clinical or echocardiographic improvement if they were eating a non-traditional diet, and the diet was changed. Although it is often thought that the only way in which cats can improve significantly after diagnosis of DCM is if the cat is taurine-deficient and receives taurine supplementation, these results suggest a possible diet-associated feline DCM that is unrelated to taurine deficiency and support the need for additional research. Although these retrospective and anecdotal survey data have important limitations, if diet-associated DCM is confirmed to occur in cats and is shown more clearly to be unrelated to taurine deficiency, it warrants further research to determine possible causative factor(s), such as deficiencies, toxicities or a combination of issues. Possible dietary causes are currently unclear but could be related to ingredients used in the formulation of some cat foods (e.g. pulses) or to other dietary factors [1,21].

Most cardiologists responding to the survey indicated that they do not believe there has been a change in numbers of feline DCM cases over the past two years. However, 15% of cardiologists did perceive an increase in the number of feline DCM cases over this time period. This discrepancy could be related to geographic differences, differences in caseload among the different cardiologists or respondent bias (i.e. only 56 of >300 cardiologists responded, so this may not have been a representative sample). In addition, the vast majority of these data come from cardiologists in the United States so may not represent findings in other countries. In addition, case numbers of feline DCM appear to be relatively low, so it may be difficult to discern even a 50–100% increase from a baseline of seeing one to two cases per year, especially for practices with multiple cardiologists.

It is noteworthy that there were no significant differences in signalment, physical examination, laboratory findings or echocardiographic measurements between diet groups other than a significantly higher body weight in cats in the high PL diet group. The relevance of this finding is unclear and does not point to a role for obesity because the body condition score was not significantly different between the diet groups, but evaluating body weight and body condition score in future, larger studies is recommended. The significant difference in body weight could also be owing to a Type I error.

Although use of ACEI in the present study was associated with significantly longer survival times on log-rank analysis, this drug was not significant on Cox proportional hazards analysis, which may be related to the relatively small sample size and the low number of cats receiving ACEI (n = 9). The use of ACEIs is often recommended for cats with DCM, especially in cats with more severe systolic dysfunction [12,15,18]; in a study of cats with non-taurine responsive DCM, all cats received an ACEI [9]. There are many reasons why an ACEI might not be instituted, including azotaemia or hypotension [22]. In some cats in the study, medications were added in a sequential approach, and cats may have died before an ACEI could be instituted. Cats that were deemed by the cardiologists to be able to tolerate an ACEI may have been less severely affected and less likely to have azotaemia or hypotension. In addition, cats in this retrospective study received a variety of medication combinations, so the effect of a single medication is difficult to identify. Similarly, the presence of CHF was associated with shorter survival times on log-rank analysis but not on Cox proportional hazards analysis. Nearly all cats (89%) had CHF at the time of diagnosis, so the ability to analyse the effect of CHF in this retrospective study was limited.

In addition to the study limitations already mentioned, there were a number of other limitations that are important to consider. Although survival time was significantly longer in cats eating high PL diets that changed diets, follow-up echocardiographic measurements were not able to be compared as they were only available for three cats at the time of writing this article. Although some cats were recently diagnosed and may still be re-evaluated, the lack of comparison of follow-up echocardiographic measurements is an important deficit, especially in consideration of four studies of dogs with diet-associated DCM in which improvements in echocardiographic measurements occurred after diet change and, in many, taurine supplementation [3–6].

In designing the study, the authors considered the lack of consensus regarding the definition of DCM in cats. Therefore, criteria from previous studies were used to construct a working definition for this study, but it is unclear whether this definition may have been too stringent, reducing the sample size unnecessarily, or too broad, resulting in the inappropriate inclusion of cats into the study and skewing the results of the data. However, our intent was to err on the side of being more stringent, and we excluded 28/67 cats (42%) submitted by cardiologists whose echocardiographic measurements did not fit within the study criteria or if key echocardiographic measurements were missing. However, despite these efforts, it is possible that some cats could have had end-stage hypertrophic cardiomyopathy. In addition, not all cardiologists performed their echocardiograms with the same approach, so there was some inconsistency in echocardiographic measurement methods among cardiologists (e.g. M-mode vs 2D measurements), which could have impacted this study. The LA:Ao was also measured using different methods which could have affected the results. The retrospective nature of the study is another important limitation, affecting the number and timing of measurements available. Future prospective studies should use consistent measurements at predetermined time points. More consistent measurement of the cardiac biomarkers, NT-proBNP and cardiac troponin I, would also be valuable in future studies.

In addition to possible respondent bias for the survey (part one), cases that were reviewed for part two of the study were solicited only from cardiologists who responded to the survey (although not every cardiologist who responded to the survey submitted cases). As a result, the sample size of the study is relatively small and potentially limits its ability to be generalised to a broader population of cats with DCM, whether diagnosed or undiagnosed, especially in countries outside the United States. Furthermore, the timeframe defined for cardiologist observations and clinical review was limited to two years; potential trends up or down in rates of feline DCM might have been better recognised and evaluated over a longer period of time. There may also have been bias in cats eating grain-free or non-traditional diets to measure taurine and recommend a diet change, so it would be useful for future studies to measure taurine in all cats and change diets in a randomised fashion (or in all cats) to minimise this bias.

Diet classification into low and high PL groups may not be the ideal approach to categorising diets. Although the cause of the apparent diet-associated DCM affecting dogs has not yet been determined, it is most often associated with diets that contain high levels of peas or lentils [1,21]. Therefore, we used these factors in defining diet categories, but further refinement of diet groups in future studies may be needed as more information becomes available. It remains unclear whether peas and the different pea fractions have similar effects. It is also impossible to determine the exact amounts of peas or lentils in the diets from the ingredient lists; some diets had a single pea ingredient in the top 10, whereas others had multiple peas and pea fractions on the ingredient list. Therefore, if high dietary levels of peas and lentils are determined to be associated with DCM in cats, further research is needed on their exact quantities in diets, rather than just a crude estimate based on the ingredient list. Another limitation was that not all cats in the study had complete diet histories available in their medical records. Owners are often unable to accurately report or remember during a veterinary visit what diets their cat is eating; this resulted in the inability to categorise three patients in either the low or high PL diet groups. For many of the cats, the information on diet in the medical records was not sufficient to be able to categorise cats into diet groups. As such, additional specific information – such as diet flavour, because flavours of an individual diet may have different formulations and affect whether it contains peas or lentils – had to be solicited by contacting primary care veterinarians or owners to be able to categorise cats. This emphasises the importance of collecting detailed information on diet at every visit or, even better, recording a photograph of the diet label and ingredient list in the medical record.

Conclusions

This study suggests that DCM is still a disease of concern in cats. When measured, taurine deficiency was uncommon, although taurine was not measured in most cats. The finding of a longer survival time for cats eating high PL diets that changed diets after diagnosis supports the hypothesis that some cases may be associated with diet. Prospective studies on this issue are recommended, including more detailed information on the prevalence of DCM in cats, and studies that evaluate the possible role of diet in feline DCM. In the meantime, it would be valuable to collect detailed dietary information on all cats with DCM, and we recommend changing the diet of all cats with DCM – regardless of the diet type – to one without peas or lentils until more information is available.

Abbreviations

2D

two-dimensional

ACEI

angiotensin-converting enzyme inhibitor

CHF

congestive heart failure

DCM

dilated cardiomyopathy

FDA

United States Food and Drug Administration

High PL diet

diet with peas or lentils in the top 10 ingredients on the ingredient list

LA:Ao

ratio of the left atrial to aortic dimensions

Low PL diet

diet with no peas or lentils in the top 10 ingredients on the ingredient list

NT-proBNP

N-terminal pro-B-type natriuretic peptide