Viewpoint articles represent the opinions of the authors and do not represent AVMA endorsement of such statements.
Introduction
Investigators recently evaluated the utility of N-terminal pro-brain natriuretic peptide (NT-proBNP) as a screening test for occult hypertrophic cardiomyopathy (HCM) in cats.1 The authors concluded that the relatively low sensitivity of the test (43%) precluded its use as a screening test in apparently healthy cats. Subsequently, 2 other authors, in a review article, supported this conclusion.2 However, this conclusion might be in error.
Screening for HCM
Two broad reasons might prompt clinicians to screen apparently healthy cats for HCM—to select breeding individuals of specific breeds that do not possess genes associated with HCM (often using a genetic test or echocardiography) and to identify individuals with occult disease—disease that is present, but in a cat that shows no historical or physical evidence of disease.
Clinicians might choose to screen cats with subclinical (occult) HCM because they appear to have a higher risk of adverse outcomes, such as congestive heart failure (CHF) or unexpected death, than healthy cats with certain interventions—anesthesia, administration of glucocorticoids or fluids—and possibly stressful situations. The data for these adverse outcomes are sparse and largely anecdotal.3,4 Nevertheless, most clinicians would like to identify cats with a substantially increased risk, so that they can either avoid situations or interventions that could trigger the adverse outcome (development of CHF, sudden death) or take precautions to reduce that risk. A screening test that would identify such “at-risk” cats would hopefully improve outcomes for such cats.
One further rationale exists for screening cats for occult HCM—to identify cats at risk of aortic (systemic) thromboembolism (ATE). The probability of a cat with subclinical HCM developing ATE at some point in time approaches 10%.5 Consequently, cardiologists generally recommend instituting antithrombotic prophylaxis (with clopidogrel) in cats with severe subclinical HCM to reduce this risk. Therefore, screening cats for moderate-to-severe occult HCM could be a reasonable goal, with the aim of decreasing these adverse outcomes.
Are all cats with HCM at increased risk of adverse events?
Although no studies have examined this directly, other studies6,7 have found that the bigger the left atrium, the more likely a cat is to develop ATE or CHF. Cats with mild HCM likely have little imminent risk of adverse outcomes of interest (perianesthetic death, CHF, or development of ATE).6 Therefore, it would make sense to use screening tests that identify cats with occult “moderate-to-severe” HCM, as these are the cats that likely carry an increased risk of adverse outcomes.
What is the prevalence of moderate to severe occult HCM in cats?
To examine the value of any screening test, we need to know the prevalence of the disease being screened for in our target population. Therefore, if we want to identify moderate-to-severe occult HCM, we need to know the prevalence of moderate-to-severe HCM in the at-large population of apparently healthy cats. Lu et al,1 who examined the diagnostic utility of the NT-proBNP assay in apparently healthy cats, provide a glimpse of what this might be, although the numbers of affected cats are small and therefore carry a large degree of imprecision in the estimate. Their study1 included 168 cats with no apparent disease and 32 with HCM, 24 of which had moderate-to-severe disease, resulting in a prevalence of 12% (10/32 cats had obstructive HCM, and therefore would not be considered occult, because they would have had murmurs). This seems rather high, given that estimates of the prevalence of subclinical HCM in the apparently healthy cat population approach 15%8–10; we would not expect most of these cats to have moderate-to-severe disease. Supporting this concern is a 1-year incidence of development of CHF or ATE in cats with subclinical HCM of 5% to 7%,5 suggesting that most cats have mild disease. Further support comes from the observation that almost no cats had moderate-to-severe disease in a large study8 of shelter cats. Finally, we cannot determine how many of the cats in the study by Lu et al1 had murmurs (these cats would no longer be considered “occult” but would be considered “subclinical”). Nevertheless, for the sake of analysis, we will assume a prevalence of 9%.
What screening test options exist for occult HCM in cats?
Currently, echocardiography serves as the reference method for diagnosing HCM in cats. Cardiologists have provided guidelines for echocardiographic screening of cats used in breeding programs.11 An abbreviated echocardiogram focusing on left atrial size (as a predictor of adverse outcomes) could suffice in many cases. However, echocardiography fails to meet most of the characteristics of screening tests—it is not readily available (it requires specific probes that most clinicians do not have), it is not cheap, it is not easily performed (it requires a certain level of expertise), and it requires expertise for correct interpretation. Similarly, radiography fails to meet many of these criteria for similar reasons; additionally, it has poor sensitivity and modest specificity for identifying left atrial enlargement.12,13 Electrocardiography also has poor sensitivity for identifying left atrial enlargement,12 is not routinely available, and requires expertise for interpretation.
The serum biomarker, NT-proBNP, offers potential as a screening test for occult HCM. It is relatively cheap, readily available, easily and safely performed, and easily interpreted. Studies14,15 have suggested relatively high sensitivity and specificity of this test in screening cats with HCM.14,15 Machen et al14 found that the test had an 84% sensitivity and 83% specificity for detecting moderate-to-severe subclinical HCM in cats referred for evaluation of a cardiac abnormality. Therefore, these cats could not be considered “occult” because they had evidence of potential cardiac disease—this is evident in the rather high number of cats with moderate-to-severe disease in that study (37/146 [25%]). Wess et al15 found high sensitivity and specificity but combined cats with CHF and cats with subclinical disease; therefore, these results do not address the problem of screening for moderate-to-severe occult HCM. However, Lu et al1 found that it has poor sensitivity (43%) but high specificity (96%) for detecting subclinical HCM of any severity, and argued that it should not be used as a screening test.
How well does NT-proBNP identify cats with moderate-to-severe occult HCM?
Lu et al1 provided some information that can help determine how well this test performs in cats with moderate-to-severe subclinical HCM. Of 32 cats with HCM, 24 had moderate-to-severe HCM. The NT-proBNP test correctly identified 11 of these 24 cats as having moderate-to-severe HCM. It also correctly identified 167 of the 200 cats as not having moderate-to-severe HCM (Table 1).
The diagnostic performance of the rapid feline N-terminal-proBrain natriuretic peptide (NT-proBNP assay in detecting moderate-to-severe occult hypertrophic cardiomyopathy (HCM) based on data from Lu et al.1
No. of cats | ||||
---|---|---|---|---|
NT-proBNP test result | Moderate-severe HCM | Mild or no HCM | Totals | |
Positive | 11 | 9 | 20 | PPV: 55% |
Negative | 13 | 167 | 180 | NPV: 93% |
Totals | 24 | 176 | 200 | |
Sensitivity: 46% | Specificity: 95% |
NPV = Negative predictive value. PPV = positive predictive value.
This resulted in a positive predictive value (PPV) of 55% (95% CI; 32% to 76%) and a negative predictive value (NPV) of 93% (95% CI; 88% to 96%). Using these results, 20 of 200 cats would have positive test results that would prompt echocardiography to confirm the diagnosis; almost half of these cats would have no disease (and therefore incur an unnecessary cost to the client). However, approximately half of the cats with moderate-to-severe HCM (13/24) would have a negative test result, and so would be mistakenly considered unaffected. These cats could then be subjected to an intervention (anesthesia, glucocorticoids, fluids) that could trigger an adverse outcome (CHF or death). Based on this observation, the authors of this paper, and, subsequently, 2 cardiologists, in a review article, suggested that NT-proBNP should not be used as a screening test, because a negative test result would lull the clinician into a false sense of security that no disease exists.1,2
Are there precedents for screening tests with similar performance characteristics?
Almost identical predicaments have existed in human medicine. Papanicolaou (Pap) smears have a 55% sensitivity and 97% specificity for detecting Grade 2 or 3 cervical epithelial neoplasia.16 These cancers have a prevalence in the “at-risk” population < 1%. As with the NT-proBNP test, Pap smears will miss detection of cervical cancers in approximately half of the affected women. The PPV of this test (the probability that a positive test is truly positive) is 16% (meaning that approximately 1 in 6 women testing positive will actually have cervical cancer), while the NPV is 99%. Nevertheless, gynecologists have screened women at risk of cervical cancer with this test for more than 50 years.
Similarly, prostate-specific antigen testing has a sensitivity of 51% and specificity of 91% for detecting high-grade prostatic neoplasia, which has a prevalence of approximately 15% of all prostate cancer diagnoses (and therefore a lower prevalence in the general population).17 This results in a PPV of 30% and an NPV of 85%. However, when the test was introduced, it was used as a screening test for decades.
With both of these screening tests, physicians accepted that the high rate of false positives would require additional testing (usually biopsies, with their own inherent risks of adverse outcomes), but considered this a reasonable trade-off for detecting diseases that responded well if identified early.
How should clinicians determine whether to test or not test for HCM?
Clinicians could view the NT-proBNP test for detecting moderate-to-severe HCM in a light similar to these well-established human screening tests. A clinician performing an NT-proBNP test on middle-aged-to-older apparently healthy cats presenting for a routine intervention (eg, a dental procedure, glucocorticoid administration) would correctly identify half of the cats of interest (those with moderate-to-severe HCM). These cats would undergo additional testing (echocardiography) to confirm the diagnosis, and then, either be spared the intervention (if elective) or be managed to try to avoid an adverse outcome (death, CHF, or ATE). Obviously, half of the affected cats would “slip through the net,” fail to have their condition detected by the test, undergo the intervention, and be exposed to the risk of an adverse outcome. Most cats, however, would test negative and be correctly identified as having no substantial disease. Finally, for several clients, their cats (false positives) would undergo additional testing (echocardiography) to refute the initial suspicion raised by a positive screening test result. Importantly, and differently from the examples of Pap smears and prostate-specific antigen testing, the “harms” of a false positive with occult HCM are economical, rather than biological.
Alternatively, clinicians can follow the suggestions proposed by Lu et al1 and Kittleson and Côté2 and not screen for HCM using this test. This strategy exposes all cats with moderate-to-severe occult HCM to an adverse outcome; all cats, whether diseased or not, will be subjected to the interventions.
Given that no alternative screening test exists, if we decide that screening cats for moderate-to-severe HCM prior to anesthesia is warranted, then identifying half of them seems to be better than identifying none of them. Even a markedly imperfect test is better than no test, provided that it does not result in an exorbitant amount of follow-up testing of cats with false positive diagnoses.
Should we bother screening cats for moderate-to-severe subclinical HCM?
As discussed earlier, clinicians should perform screening tests if they allow early intervention to delay disease progression, or a change in management to avoid or minimize adverse outcomes. Additionally, the adverse outcome must be important enough, or the prevalence high enough, to warrant the expense of testing a population of “at-risk” individuals.
Therefore, does moderate-to-severe disease confer a high risk of adverse perianesthetic outcomes (development of CHF or sudden death)? This is difficult to answer—anecdotally, clinicians are concerned that IV fluid therapy or even just the stress of anesthesia can precipitate the development of CHF. A study by Rush et al4 supported this idea; they found that half of the cats with CHF had a potential precipitating event, such as fluid administration, anesthesia, or glucocorticoid administration, shortly prior to the CHF diagnosis.
Another way to approach this is to look at the perianesthetic death of apparently healthy cats undergoing anesthesia. Two studies18,19 have examined this in detail; these studies found that the perioperative mortality rate in apparently healthy cats fell somewhere between 0.05% and 0.1%; in other words, a death every 1,000 to 2,000 anesthetic procedures. None of the cats in these studies underwent screening for occult HCM. If occult moderate-to-severe HCM has a prevalence of 10%, as suggested by the study of Lu et al,1 we might expect this mortality rate to be substantially higher in cats that are not screened for HCM. Even if we estimate that 10% of cats with moderate-to-severe HCM (1% of all cats) suffer an adverse outcome, the perioperative mortality rate would be expected to be substantially higher than reported. However, Brodbelt and colleagues18 did find that both increasing age and administration of fluids during anesthesia increased the risk of perianesthetic death. While they did not document heart disease in these cats, occult cardiomyopathy could potentially explain such an association.
Of course, cats with moderate-to-severe HCM might not die during or after anesthesia but might develop CHF, an adverse outcome that would not show up in the studies by Brodbelt et al18 and Levy et al.19 No data exist examining the risk of cats with HCM developing CHF after anesthesia. Nevertheless, the risk appears to be relatively low.
What about identifying cats with moderate-to-severe occult HCM to institute antithrombotic treatment, or prior to administration of glucocorticoids? The probability of these outcomes in cats with HCM remains low, even in cats with moderate-to-severe disease. Therefore, it currently makes little sense to routinely screen cats with the hope of identifying cats likely to suffer ATE.
Therefore, it might be that clinicians overestimate the risks of adverse events in cats with moderate-to-severe subclinical HCM subjected to triggering interventions. If this is true, then screening for such cats is medically unwarranted (most cats would test negative, and those that were identified with disease would have likely no adverse outcomes following the intervention).
However, intangible factors influence a clinician’s decision to perform diagnostic testing; concerns of chagrin or regret if a patient has an adverse event that could have been mitigated or prevented,20,21 or fear of retribution by a client whose pet suffered the adverse event (either on social media or as formal malpractice complaints). Therefore, even if certain screening tests appear unnecessary or imperfect, these factors can sway clinicians to perform them so that, in the event of an adverse event, they can claim, “At least we ran the test to try to identify the disease.” That argument is much easier to stomach than “If only we had run the test, we might have detected the disease.” Or trying to answer the distraught owner’s question: “Why did you not test for the disease if you knew a test existed?”
Conclusions
This article raises 2 points for consideration: should cats be screened for HCM to try to reduce, avoid or prevent adverse outcomes, and if so, what screening test should clinicians use? There is no correct answer to the first question, although the current understanding of the risks of these adverse outcomes is rudimentary, and relies largely on extrapolation of circumstantial evidence. The answer to the second question appears more clear: if a clinician decides that screening is worthwhile, the only test, albeit quite imperfect, that meets most of the requirements of a screening test is the NT-proBNP assay. The test brings to mind the adage “Perfect is the enemy of the good.” Or, as the Bard put it most eloquently, “Were it not sinful then, striving to mend, to mar the subject that before was well?”
Acknowledgments
No third-party funding or support was received in connection with this study or the writing or publication of the manuscript. The author declares that there are no conflicts of interest.
References
- 1. ↑
Lu TL, Côté E, Kuo YW, et al. Point-of-care N-terminal pro B-type natriuretic peptide assay to screen apparently healthy cats for cardiac disease in general practice. J Vet Intern Med. 2021;35(4):1663–1672. doi:10.1111/JVIM.16156
- 2. ↑
Kittleson MD, Côté E. The feline cardiomyopathies: 2. hypertrophic cardiomyopathy. J Feline Med Surg. 2021;23(11):1028–1051. doi:10.1177/1098612X211020162
- 3. ↑
Smith SA, Tobias AH, Fine DM, Jacob KA, Ployngam T. Corticosteroid-associated congestive heart failure in 12 cats. Int J Appl Res Vet Med. 2004;2(3):159–170.
- 4. ↑
Rush JE, Freeman LM, Fenollosa NK, Brown DJ. Population and survival characteristics of cats with hypertrophic cardiomyopathy: 260 cases (1990–1999). J Am Vet Med Assoc. 2002;220(2):202–207. doi:10.2460/javma.2002.220.202
- 5. ↑
Fox PR, Keene BW, Lamb K, et al. International collaborative study to assess cardiovascular risk and evaluate long-term health in cats with preclinical hypertrophic cardiomyopathy and apparently healthy cats: The REVEAL Study. J Vet Intern Med. 2018;32(3):930–943. doi:10.1111/JVIM.15122
- 6. ↑
Payne JRR, Borgeat K, Connolly DJJ, et al. Prognostic Indicators in Cats with Hypertrophic Cardiomyopathy. J Vet Intern Med. 2013;27(6):1427–1436. doi:10.1111/jvim.12215
- 7. ↑
Payne JR, Borgeat K, Brodbelt DC, Connolly DJ, Fuentes VL. Risk factors associated with sudden death vs. congestive heart failure or arterial thromboembolism in cats with hypertrophic cardiomyopathy. J Vet Cardiol. 2015;17(suppl 1):S318–S328. doi:10.1016/j.jvc.2015.09.008
- 8. ↑
Payne JR, Brodbelt DC, Luis Fuentes V. Cardiomyopathy prevalence in 780 apparently healthy cats in rehoming centres (the CatScan study). J Vet Cardiol. 2015;17(suppl 1):S244–S257. doi:10.1016/j.jvc.2015.03.008
- 9.
Paige CF, Abbott JA, Elvinger F, Pyle RL. Prevalence of cardiomyopathy in apparently healthy cats. J Am Vet Med Assoc. 2009;234(11):1398–1403. doi:10.2460/javma.234.11.1398
- 10. ↑
Côté E, Manning AM, Emerson D, Laste NJ, Malakoff RL, Harpster NK. Assessment of the prevalence of heart murmurs in overtly healthy cats. J Am Vet Med Assoc. 2004;225(3):384–388. doi:10.2460/javma.2004.225.384
- 11. ↑
Häggström J, Luis Fuentes V, Wess G. Screening for hypertrophic cardiomyopathy in cats. J Vet Cardiol. 2015;17(suppl 1):S134–S149. doi:10.1016/j.jvc.2015.07.003
- 12. ↑
Schober KE, Maerz I, Ludewig E, Stern JA. Diagnostic accuracy of electrocardiography and thoracic radiography in the assessment of left atrial size in cats: comparison with transthoracic 2-dimensional echocardiography. J Vet Intern Med. 2007;21(4):709–718. doi:10.1892/0891-6640(2007)21[709:daoeat]2.0.co;2
- 13. ↑
Guglielmini C, Diana A. Thoracic radiography in the cat: identification of cardiomegaly and congestive heart failure. J Vet Cardiol. 2015;17(suppl 1):S87–S101. doi:10.1016/j.jvc.2015.03.005
- 14. ↑
Machen MC, Oyama MA, Gordon SG, et al. Multi-centered investigation of a point-of-care NT-proBNP ELISA assay to detect moderate to severe occult (preclinical) feline heart disease in cats referred for cardiac evaluation. J Vet Cardiol. 2014;16(4):245–255. doi:10.1016/j.jvc.2014.09.002
- 15. ↑
Wess G, Daisenberger P, Mahling M, Hirschberger J, Hartmann K. Utility of measuring plasma N-terminal pro-brain natriuretic peptide in detecting hypertrophic cardiomyopathy and differentiating grades of severity in cats. Vet Clin Pathol. 2011;40(2):237–244. doi:10.1111/j.1939-165X.2011.00305.x
- 16. ↑
Mayrand M-H, Duarte-Franco E, Rodrigues I, et al. Human papillomavirus DNA versus Papanicolaou screening tests for cervical cancer. N Engl J Med. 2007;357(16):1579–1588. doi:10.1056/NEJMoa071430
- 17. ↑
Adhyam M, Gupta AK. A review on the clinical utility of PSA in cancer prostate. Indian J Surg Oncol. 2012;3(2):120–129. doi:10.1007/s13193-012-0142-6
- 18. ↑
Brodbelt DC, Pfeiffer DU, Young LE, Wood JLN. Risk factors for anaesthetic-related death in cats: results from the confidential enquiry into perioperative small animal fatalities (CEPSAF). Br J Anaesth. 2007;99(5):617–623. doi:10.1093/bja/aem229
- 19. ↑
Levy JK, Bard KM, Tucker SJ, Diskant PD, Dingman PA. Perioperative mortality in cats and dogs undergoing spay or castration at a high-volume clinic. Vet J. 2017;224:11–15. doi:10.1016/j.tvjl.2017.05.013
- 20. ↑
Feinstein AR. The “chagrin factor” and qualitative decision analysis. Arch Intern Med. 1985;145(7):1257–1259.
- 21. ↑
Müller BS, Donner-Banzhoff N, Beyer M, Haasenritter J, Müller A, Seifart C. Regret among primary care physicians: a survey of diagnostic decisions. BMC Fam Pract. 2020;21:(1);57. doi:10.1186/s12875-020-01125-w