Is thoracic ultrasound more efficient than the Wisconsin calf scoring system for the detection of pneumonia in calves?

There is some evidence to suggest that thoracic ultrasonography could be more accurate than the Wisconsin calf scoring system for detecting pneumonia in calves. However, the limitations of the studies assessed in this evidence evaluation mean that it is difficult to draw a definitive conclusion. Although it was not the focus of this evidence evaluation, there may be value in investigating the combined use of thoracic ultrasonography and the Wisconsin calf scoring system in certain circumstances to maximise the identification of cases. At the end of a routine herd health visit on a large dairy farm, the farmer mentions that he is planning to start a separate business buying in and rearing preweaned dairy cross calves for beef and asks you to cast an eye over his new calf-rearing shed. After discussing building design and vaccination protocols, the farmer notes that a neighbour has started to use thoracic ultrasonography (TUS) to check their calves for pneumonia. He asks you whether this would be better at detecting pneumonia in groups of calves than the Wisconsin calf scoring (WCS) system that he currently uses. You wonder if there is any evidence comparing the efficiency of TUS and WCS for detecting pneumonia in calves. In [calves at risk of pneumonia], does [thoracic ultrasonography compared with the Wisconsin calf scoring system] lead to [more efficient disease detection]? The search strategy can be viewed at https://bestbetsforvets.org/bet/575, and is also available as a supplement to this article on Vet Record's website at https://bvajournals.onlinelibrary.wiley.com/toc/20427670/2021/189/2 One hundred and ninety-one papers were found in the Medline search. One hundred and forty-six were excluded because they did not answer the question. Twenty-two were excluded because they were not written in English. Nineteen were excluded because they were review articles, in vitro research or conference proceedings. In total, four relevant papers were obtained. One hundred and twenty-four papers were found in the CAB search. Seventy were excluded because they did not answer the question. Twenty-two were excluded because they were not written in English. Twenty-eight were excluded because they were review articles, in vitro research or conference proceedings. In total, four relevant papers were obtained. Overall, four relevant papers were found. Search last performed: 13 June 2021 Paper 1: Bayesian estimation of the accuracy of the calf respiratory scoring chart and ultrasonography for the diagnosis of bovine respiratory disease in preweaned dairy calves1 Patient group: Datasets from two North American field studies in which preweaned Holstein calves were assessed for pneumonia were evaluated. In both studies, the calves were simultaneously assessed using TUS and the WCS system. The first study2 involved 85 calves from 13 dairy farms in Quebec, Canada. These calves were anticipated to have a high prevalence of pneumonia; however, there is little baseline information given. The second study3 involved 106 calves from six dairy farms in New York State, USA, with no history of calf pneumonia. Study type: This paper is a short communication. However, both studies referenced within it are cohort studies. Outcomes: In both studies, the case definition for pneumonia was a score of at least 5 on the WCS scale or a parenchymal consolidation depth of at least 1 cm observed on the thoracic sonogram. Based on these case definitions, four Bayesian latent class models were used to calculate the prevalence of pneumonia in the two study populations, the sensitivity and specificity of the two diagnostic tests and the conditional covariance between the tests. Key results: The agreement between the two tests was slight to fair for both the New York State population (κ=0.133, 95 per cent confidence interval [CI] −0.551–0.321) and the Quebec population (κ=0.292, 95 per cent CI 0.030–0.554). For model 1, where it was assumed the sensitivity and specificity of two tests were the same for both study populations, the median sensitivity and specificity of the WCS system were estimated to be 62.4 per cent (95 per cent credibility interval [CreI] 47.9–75.8) and 74.1 per cent (95 per cent CreI 64.9–82.8), respectively. Meanwhile, the median sensitivity and specificity of TUS were estimated to be 79.4 per cent (95 per cent CreI 66.4–90.9) and 93.9 per cent (95 per cent CreI 88.0–97.6), respectively. However, the credibility intervals of the sensitivity estimates were relatively wide for both tests. All estimates were broadly similar across all four models when prevalence priors were altered (variation of less than 5 per cent between models), with the exception of TUS sensitivity. For this estimate, the median across all models varied between 70.3 and 79.4 per cent. Again, the credibility intervals of the sensitivity estimates were relatively wide for both tests across all four models. Study weaknesses: Although the aim stated in the introduction was to estimate the accuracy of both TUS and WCS for the diagnosis of pneumonia in preweaned dairy calves, the conclusion pertains to the characteristics of the WCS system only. No reasoning was given for including data from these two studies particularly, and the sample is unlikely to be representative of the target population as data from one of the studies was based on convenience sampling. In addition, the small number of animals included in each study limits the extent to which the findings can be generalised. Many assumptions were made about the priors that were input into the models due to the lack of available baseline data. Although there was some description of the rationale behind using informative/non-informative priors for a parameter, alongside how estimates were obtained from experts for informative priors, more information about the priors and assumptions would be useful to understand and interpret the results fully. The short communication format of this paper may be responsible for the scant detail provided, especially in relation to the diagnostic testing approach used, the amount of basic data provided from the two original studies and the explanation of the cut offs used for accuracy. Paper 2: Bayesian estimation of sensitivity and specificity of systematic thoracic ultrasound exam for diagnosis of bovine respiratory disease in preweaned calves4 Patient group: Data from two prospective studies were included in the analysis. The first study5 was a randomised clinical trial assessing the antimicrobial efficacy of tildipirosin. It involved 209 veal calves, from a single farm in Quebec, that were approximately seven days old at enrolment. The second study6 involved 301 dairy calves from 39 herds visited by the ambulatory clinic of the University of Montreal in 2015. These calves were approximately 38 days old at enrolment and were part of a larger study validating the Californian clinical scoring (CCS) system – which is the same as WCS but with fever also included as a clinical sign –by comparing it with TUS.7 Study type: The studies referenced within this study were a randomised controlled trial5 and a cross-sectional study.6 Outcomes: Estimates of sensitivity, specificity and positive and negative likelihood ratio were calculated for three different TUS case definitions (depth of consolidation ≥0 cm, ≥1 cm and ≥3 cm) using a Bayesian latent class model. The same estimates were also calculated for four TUS scan sites (right lung lobe cranial to the heart, right and left lung lobes caudal to the heart, right cranial and right and left lung lobes caudal to the heart) using the same model. Four Bayesian latent class models were then used to calculate the prevalence of pneumonia in the two study populations, the sensitivity and specificity of the two diagnostic tests, the conditional covariance between the tests and the positive and negative predictive values for TUS. These outcomes were calculated using the case definitions of a score of at least 5 on the CCS scale or a parenchymal consolidation depth of at least 3 cm observed by TUS. Key results: In the initial model, the median sensitivity and specificity of the CCS scale were 69 per cent (95 per cent CreI 40–97) and 95 per cent (95 per cent CreI 92–97), respectively. Meanwhile, the median sensitivity and specificity of TUS were 89 per cent (95 per cent CreI 55–100) and 95 per cent (95 per cent CreI 92–98). There was little effect on TUS sensitivity and specificity estimates when dependency between tests was considered (model 2). However, CCS sensitivity became much lower (0.48 per cent), and there was more evidence of dependency when calves did not have active pneumonia. The sensitivity and specificity estimates did not vary much when considering the farming system (veal versus dairy, models 3 and 4). Study weaknesses: No reasoning was given for including data from these two studies particularly, and the sample is unlikely to be representative of the target population as data for one of the studies were obtained from just one farm. In addition, no sample size calculation was included, so the analysis may be underpowered. It is unclear whether TUS and CCS were independently performed, as combinations of the same operators performed both tests. Furthermore, the origin of the CCS data for the dairy population was difficult to identify without referring to the larger study these calves were part of.7 Information concerning the definition and rationale of the prior information provided was lacking. Consequently, there was not enough information to fully understand and interpret the priors, which is vital for accurate assessment of the results. The low prevalence of pneumonia in both study populations studied would also have affected the precision of the estimates, although this is discussed as a limitation. Paper 3: Growth performance and haematological changes of weaned beef calves diagnosed with respiratory disease using respiratory scoring and thoracic ultrasonography8 Patient group: A total of 153 weaned male (n=79) and female (n=74) calves (average age 209 days) were purchased through 10 auction markets and housed at an Irish research centre. Forty-seven of these calves were sired by Aberdeen Angus and Hereford breeds, and 106 were sired by Charolais and Limousin breeds. Study type: Cohort study. Outcomes: Calves were assessed for pneumonia on days 0, 7, 14 and 28 using both TUS and WSC. TUS was scored on a three-point scale: 0=no lung consolidation, 1=at least one comet tail artefact observed, 2=1 cm or more of lung consolidation. For WCS, each clinical sign included in the scoring system was assigned a score from 0 (normal) to 3 (very abnormal), and animals with a combined score of 5 or more were classified as positive. The distribution of the WCS and TUS scores and the prevalence of WCS-positive animals and those with a TUS score of 2 were then calculated for days 0, 7, 14 and 28. Pearson's correlation coefficient was also used to assess the relationship between the WCS clinical signs and lung consolidation on days 7, 14 and 28. Key results: Based on WCS, 35 per cent of these calves were diagnosed with pneumonia. Of these cases, 50 per cent were detected within seven days of arrival at the research centre, while 81 per cent were detected within the first 14 days. However, TUS detected no lung consolidation in 56 per cent of the WCS-positive calves. A greater number of pneumonia cases were detected when both TUS and WCS were used together, with TUS detected lung consolidation in 28 per cent of the calves classified as WCS-negative. However, there appeared to be no correlation between the percentages of calves with clinical respiratory signs and lung consolidation at days 14 and 28 (P>0.05). Study weaknesses: All calves were assessed by the same trained research vet; therefore, the assumption is that the WCS and TUS assessments were not independent or blinded. Furthermore, cranial lung lobes could not be scanned in this study due to the greater age of the calves, and resultant detection of pneumonia may have been limited by this. The incidence of calves with clinical signs of pneumonia was 35 per cent. However, animals were vaccinated against major respiratory pathogens on arrival at the research centre, which may have affected disease incidence and made the associations drawn between clinical signs and TUS scores less credible. In addition, it is unknown how representative of beef production systems in Ireland these animals might be, which makes extrapolating the findings difficult. Paper 4: Association between clinical respiratory signs, lung lesions detected by thoracic ultrasonography and growth performance in preweaned dairy calves9 Patient group: A total of 28 male Holstein Friesian-Aberdeen Angus cross calves (average age 24 days) and 25 male Holstein Friesian calves (average age 21 days) were purchased from 13 dairy farms and housed at a research centre in Ireland. Study type: Cohort study. Outcomes: Calves were assessed for pneumonia on days 0, 7 and 14 and 28 using both TUS and WSC. Between days 14 and 30, 33 of the calves were examined using TUS only. TUS was scored on a four-point scale: 1=no lung lesions, 2=a lung lesion less than 2 cm2, 3=a lung lesion of 2 cm2 or larger, 4=consolidated lung lobe/emphysema. For WCS, each clinical sign included in the scoring system was assigned a score from 0 (normal) to 3 (very abnormal), and animals with a combined score of 5 or more were classified as positive. The distribution and frequency of the scores for each clinical sign and the TUS scores were calculated from day 0 to day 14 (clinical scores and TUS scores) and from day 14 to day 30 (TUS scores only). Spearman's rank correlation was then performed to assess the relationship between each clinical sign and the TUS score on days 0, 7 and 14. Key results: TUS detected lung lesions in 64 per cent of the calves between purchase and weaning on day 53, while 43 per cent were classified as WCS-positive during this time. Of the calves with lung lesions, 44 per cent were WCS-negative. Furthermore, 61 per cent of the WCS-positive calves had lung lesions before they began to display clinical signs. Only a weak to moderate correlation was detected between TUS and clinical signs (rsp=0.40, P<0.05). Study weaknesses: The number of animals included in this study was based on a power calculation for a different study looking at a different outcome, but the degree of equivalence between the two studies is unclear. It is also unknown how representative these animals might be of dairy farms in Ireland. Assessments for the first 14 days of the trial were performed by the same people and are, therefore, unlikely to be independent or blinded. However, it is unclear who performed the later assessments. Furthermore, all calves were assessed by using WCS and TUS on days 0, 7 and 14. However, 33 of the 53 calves underwent TUS between days 14 and 30, with no explanation provided as to why. Individual clinical signs used in the WCS system, not overall WCS scores, were used to compare WCS and TUS for some of the data analysis. Assessing clinical signs individually is not a previously validated system that is proven to be correlated with pneumonia; hence the two tests have not been directly compared in the analysis. However, this, and other limitations of the study, were not discussed by the authors. The question for this evidence evaluation was formulated to investigate the performance of the two diagnostic tests in calves of all age groups. However, after completing the evaluation, it has become apparent that TUS and clinical scoring characteristics may vary between age groups due to obvious differences in size and the resultant impact on ultrasound scanning techniques. Further work to investigate this would be beneficial. The methods of diagnostic testing used in these studies were often not explicit. Although every study appraised referenced the practical ultrasonography approach used, there is no standardisation between studies concerning the technique used or the classification of a ‘positive’ lesion. As such, it may be challenging to extrapolate the outcomes of these studies more widely. The aetiology of calf pneumonia was not investigated in any of the studies assessed here. It is possible that different aetiologies may lead to differing thoracic lesion sites and features, which would affect the performance of TUS in studies where only a restricted site on the thorax was scanned. Aetiology may also differ between countries, again making the extrapolation of outcomes more challenging. Additionally, the relationship between TUS and WCS does not appear to be well correlated. This may be because TUS is more efficient at detecting disease or because clinical signs are associated with the early stages of pneumonia while lung lesions are associated with later stages. Finally, the prevalence of pneumonia in most of the study populations was low, and no tangible power calculations were presented to support the sample size in any of the studies. This ultimately makes it difficult to draw definitive conclusions. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article. The ‘Evaluating The Evidence’ section of Vet Record aims to answer specific clinical questions using a systematic approach to identify and succinctly summarise the relevant evidence from the scientific literature. The shortcomings of this evidence are also taken into account, thereby enabling vets to incorporate the best available evidence from the literature when making clinical decisions. Please contact us at vet.research@bvajournals.com if you have an article you would like us to consider for publication in this section.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.