Non-enterotoxigenic (containing neither fimbrial nor toxin genes) E. coli was a frequent finding in both diarrhoeic and non-diarrhoeic piglets, whereas only one enterotoxigenic isolate was detected. Hence, ETEC did not seem to play any pathogenic role in relation to the investigated outbreaks of diarrhoea. Other studies have indicated that attaching and effacing E. coli (AEEC), carrying neither fimbrial nor toxin genes, are able to induce diarrhoea in newborn piglets  and to induce villous atrophy . The prevalence of AEEC in the present study is currently being investigated.
CPC was cultured in four piglets of the study. Due to the low prevalence the significance of this bacterium in relation to the investigated outbreaks is probably minimal. The significance of CPA in relation to diarrhoea in neonatal piglets is controversial, since it has been concomitantly recognized as part of the normal intestinal flora and as a potential intestinal pathogen [11, 12]. In this study we found a significantly higher prevalence of CPA in non-diarrhoeic vs. diarrhoeic piglets. Most likely, this finding merely reflects the intact intestinal flora within the non-diarrhoeic piglets. CD has been reported in cases of neonatal diarrhoea in piglets [5, 12]. However, in this study, this bacterium was only detected in two piglets and the characteristic histopathological lesions previously reported to be associated with CD infections  were not seen. Accordingly, CD does not seem to be associated with the investigated outbreaks.
The scarcity of known pathogens in the outbreaks of the study supports the hypothesis that the investigated herds experienced diarrhoea of unknown aetiology. Therefore the outbreaks may be representative of the new syndrome NNPDS.
A poor body condition and dehydration were rather prevalent findings in diarrhoeic piglets in this study. However, unless very pronounced, these features are not characteristic for specific diarrhoeic syndromes, since they merely reflect the loss of nutrients and water associated with any diarrhoeic condition. Milk-filled stomachs, in contrast, seem to be a characteristic finding associated with this syndrome. Since neonatal diarrhoea is commonly associated with malabsorption caused by starvation, the filled stomachs seen in 100% of diarrhoeic piglets in this study are interesting findings which clearly differentiate this syndrome from outbreaks of neonatal diarrhoea related to starvation. However, since the vast majority (80%) of piglets in this study were diarrhoeic for two or three days only, we do not have information on the contents of stomachs at later stages of disease. Obviously, one would expect long lasting diarrhoea to keep piglets from suckling due to malaise, and therefore this criterion is probably only valid at early stages of disease.
Intestinal flaccidity was the most prominent and consistent gross lesion. Flaccidity of intestines is seen in different conditions, Postweaning Multisystemic Wasting Syndrome (PMWS)  and diet-induced malabsorption being the most obvious examples. Liquid contents in colon are expected in all diarrhoeic conditions and are therefore not considered diagnostic to any specific syndrome. Somewhat surprising, half of the diarrhoeic piglets did not have liquid content in colon at necropsy, which probably reflects that these piglets were in the recovery phase of disease. If so, this potentially implies a diagnostic problem due to less pronounced lesions and decreased excretion of infectious agents at this phase. However, since the clinical course of diarrhoea turned out to be very short (as evidenced by 80% of the selected piglets being diarrhoeic for only two or three days) it would not have been practically feasible to avoid selection of piglets in recovery.
The most consistent and predominant histological lesion observed in diarrhoeic piglets was villous atrophy (seen in 63% of diarrhoeic vs. 12% of non-diarrhoeal piglets). Villous atrophy is a very common finding in diarrhoeic conditions  and in this study, the atrophy was neither associated with infection by well-known pathogens nor malnutrition. The strong association between villous atrophy and grossly visible intestinal flaccidity indicates that decreased mucosal thickness is reflected grossly as a thin-walled, atonic intestine. Epithelial lesions in the large intestine also seemed to be consistently associated with diarrhoea in this study (seen in 33% of diarrhoeic vs. 11% of non-diarrhoeic piglets), but due to the low prevalence, these lesions do not seem to be relevant to include in a case definition.
Overall, the present study suffers from lack of comparable piglets from non-NNPDS-affected herds in order to correctly classify findings as typical or diagnostic of NNPDS. Moreover, the selection of study herds posed some difficulties since the selection was basically based on a high prevalence of diarrhoea and absence of agents (in five piglets). Obviously, potential misclassification of herds is an issue to consider – though hard to address or control at this stage of investigation.
To our knowledge, this is the first study investigating outbreaks of diarrhoea in herds suspected to suffer from NNPDS. The aetiology behind these outbreaks was either undetected pathogens or non-infectious factors. Practical experience indicates that eg. high levels of protein in sow feed can lead to diarrhoea in neonatal pigs. However, all of the investigated herds used restricted levels of protein in sow feed, and had previously tried minimizing protein content with no preventive effect. As previously underlined, the diarrhoea seemed unrelated to postnatal starvation, bot intrauterine events may have affected the normal development of intestinal absorptive capacity. Thus, the villous atrophy seen in the study may reflect prenatal under-development of villi.
The unspecific nature of intestinal lesions seen in this study underlines the complexity of intestinal pathology in neonatal pigs. Interestingly, even early studies from the seventies and eighties concluded that gross lesions seem similar and unspecific irrespective of the underlying aetiology in this age group of piglets [16, 17]. Moreover, in the study from 1975 no pathogens were detected in as many as 32% of fatal neonatal gastroenteropathies. Thus, the existence of neonatal diarrhoa with unspecific lesions and without known pathogens is not a new phenomenon. It appears, however, that the clinical picture in the herds is new. At this point we do not know whether this clinical picture is a more severe manifestation of a syndrome already present but not recognized back in the seventies, or if we are experiencing a truly new syndrome.
From a practical point of view, obviously the most urgent issue is to recognize the aetiology behind the current problems. This study disclaims associations with established agents, but yet unestablished agents or shifts in intestinal bacteral population dynamics may play a role. Therefore, culture-independent methods like metagenomics and high-throughput qPCR may be rewarding in the future investigation.