The Swine Health and Production author guidelines for original research state, among other requirements, that the conclusions of the study should be supported by the data. This is obviously an appropriate and important part of the peer-review process. However, an objective review of a recently published article raises questions about the rigor and consistency of application of this requirement in our journal.
The article, "Evaluating a natural outbreak of porcine proliferative enteropathy and treatment with tylosin in the grow finish phase," (SHAP. 1998;6(2):67-72) draws the conclusion that an intervention "was effective in treating PPE" when the data actually supports the opposite conclusion. Specifically, there were no significant differences (P>.05) between treated and untreated pigs in prevalence of PPE-like gross lesions (ileal thickening) at slaughter, the only parameter approaching a PPE-suggestive effect documented in the study (26% lesion prevalence in both groups). Nor were there any significant differences between groups for ADG, ADFI, or F:G for any period in the study. There were significant differences between groups for average clinical impression scores (CIS); however, CIS are subjective and, like productivity measures, are not specific for PPE or any other condition. Therefore an objective interpretation of the data supports a conclusion that the intervention was not effective for treating PPE or improving the rate or efficiency of growth in the presence of PPE. Simply ruling out two possible confounding infections (Serpulina and Salmonella) doesn't imply PPE specificity, particularly when the lone PPE-suggestive effect (prevalence of ileal thickening) showed no treatment effect. Forming questionable conclusions based upon 'tendencies,' and speculation that increasing the sample size may possibly have facilitated detection of significant differences, seems inappropriate in a peer-reviewed scientific journal.
The pathognomonic lesion for PPE is intestinal crypt cell hyperplasia with the presence of intracellular curved bacteria. Histopathologic examination with hematoxylin/eosin and silver stains (Warthin Starry) easily confirms PPE lesion specificity with much greater sensitivity than gross lesion assessment. The study design would have been strengthened considerably by utilizing this routine, inexpensive, and readily available technique on slaughter samples. In fact, the authors used these very techniques as one measure of confirmation of Lawsonia infection at the start of the study. If, when designing the study, the authors felt that the PigMON ileal palpation technique at slaughter had inadequate sensitivity and specificity, why did they use it rather than the more sensitive and specific standard histopathological techniques?
We look forward to receiving many more issues of Swine Health and Production. Increasing the stringency of the manuscript requirements and peer review process can only strengthen this already outstanding journal.
--Don Walter, DVM, Boehringer Ingelheim-NOBL;
Kent Schwartz, DVM, MS, TEAM Associates
Note from the Editor: We appreciate these comments from Drs. Walter and Schwartz. As you know, every article in Swine Health and Production is reviewed for scientific rigor and relevance prior to being published. The review process always improves the final product, but as scientists, we must still read with a critical eye. Drs. Walter and Schwartz raise issues of statistical significance, subjective evaluation, and interpretation of results. These are extremely important questions. An excellent response from the authors is below.
The authors reply:
Thank you for your interest in our research. The article you mentioned summarizes the findings of a single study from a natural outbreak of PPE and subsequent treatment with Tylan. It is not meant to be a complete summary of all research findings regarding Tylan and PPE.
It is correct that not all parameters evaluated indicated a statistical significance at P<.05; however, the data does support the conclusion of efficacy. Although a P<.05 is often used to indicate statistical significance, attaining a value between .05 and .10 does not indicate a lack of a difference between treatments. Statistical significance can be influenced by many factors and is often largely influenced by sample size. The statistical analysis used in this study is based on pen as the experimental unit with four pens per treatment. We reported actual P values to allow readers to decide the significance of this data, keeping in mind that the statistics are an aid in interpreting data. In field outbreaks, it is not always possible to have a large enough sample size to detect a difference at P<.05; however, these studies often generate preliminary results that merit further investigation utilizing several trials of a similar protocol and pooled trial analysis. This study is an indication of Tylan's efficacy in controlling PPE.
"Subjective" measures of disease like clinical impression scores are frequently utilized as an indicator of disease severity and the impact of intervention strategies in an outbreak situation. Evaluating changes in clinical disease over time gives important information about the pattern of pig response during the treatment period. Single timepoint measures taken at the beginning and end of an outbreak do not clearly reflect the clinical course of the disease relative to severity and response over time.
The authors agree that a diagnosis of ileitis can be confused with many things. In addition to ileitis, the differential diagnosis may include primary bacterial agents like Serpulina and Salmonella species, secondary bacteria such as E. coli or Clostridium or even viruses such as transmissible gastroenteritis virus (TGEV). In this case the diagnosis of porcine proliferative enteropathy (PPE) was well substantiated based on gross necropsy, histopathologic confirmation using H & E and Warthin-Starry staining, and the ruling out of other infectious agents. Further substantiation of PPE was achieved with positive results from fecal samples on dot blot hybridization for Lawsonia intracellularis done at the beginning of the trial. Other poorly defined conditions such as hemorrhagic bowel syndrome, gastric ulcers, and intestinal torsion were ruled out based upon gross necropsy and subsequent histopathologic examination confirming the PPE.
We agree that confirmation of ileal thickening at slaughter with Warthin-Starry staining would be a more specific test; however, we do not agree that it would have been a more sensitive indicator. Jones, et al.,1 indicated macroscopic thickening of the ileum at slaughter was 100% sensitive and 58%-89% specific for PPE. Utilizing silver staining could have added to the specificity. Epidemiologic studies have indicated that intestinal samples may be positive on Warthin-Starry at 3 weeks post exposure; however, at 10 weeks after exposure to Lawsonia intracellularis, thickening without the presence of intracellular organisms may be expected.2 In this study, pigs at slaughter would not have been expected to have the infective agent still present intracellularly and further confirmation of the diagnosis was not warranted. The presence of intestinal thickening at slaughter was not expected and we considered this finding worth reporting.
We stand by the conclusions and implications of this study and appreciate your interest in our results.
--Melissa F. Veenhuizen, DVM MS; Daniel H. Mowrey, PhD;
Gregory M. Moore, DVM; Lee E. Watkins , PhD
1. Jones et al. Comparison of techniques for diagnosis of proliferative enteritis of swine. AJVR. 1993; 54(12):1890-1985.
2. Smith SH, McOrist S. Development of persistent intestinal infection and excretion of Lawsonia intracellularis by piglets. Res Vet Sci. 1997; 62(1):6-10.
IPVS '98--An update
Dear Editor: As we get closer to July and the 15th IPVS Congress, I thought I should let you know how things are progressing. We sent out our first mailing of 7000 brochure packs last year, and posted the final announcement the last week in January. AASP members should have their copies by now. We have already received registrations, and hope to welcome about 1500 delegates from around the world.
At the last count, we've received over 850 abstract submissions. Stan Done, the Chairman of the Scientific Subcommittee, is still working on them to construct the final program of presentations. So far, it looks as though there will be over 300 oral presentations, 450 posters, and 80 papers in the Clinical Sessions. There is plenty of space in the Convention Centre, so posters are virtually guaranteed a place.
Submissions have come everywhere from Iceland to New Zealand, Japan to Brazil, and all points between! The Core Program of 50 invited review papers is finalized, and summarized in the final announcement leaflet--we may get it all on the website, with luck. We have papers on all the diseases in the news: circovirus/MSWS, pneumovirus/PRSV, PRRS, CSF, and the rest--plus lots of other good stuff. We keep the website updated regularly--it's at www.ipvs98.demon.co.uk.
The social program looks good too--Hanbury Hall is the venue for what promises to be a wonderful Farewell Banquet, while the Welcome Reception will have music from an English folk band that includes a guitar-playing expert in environment control.
Don't forget: you get FOUR days of good science (plus the OIE satellite meeting on CSF to follow), THREE volumes of proceedings (one being a stand-alone review text), TWO great social events, and ONE GOOD TIME.
We look forward to seeing you in Birmingham in July!
--Dr. John Wilkinson for IPVS '98