A retrospective study of clinical and laboratory features and treatment on cats highly suspected of feline infectious peritonitis in Wuhan, China

Feline infectious peritonitis (FIP) is a systemic, potentially fatal viral disease. The objectives of this study were to review clinical and laboratory features and treatment of cats highly suspected of FIP in Wuhan, China. The clinical records of 127 cats highly suspected of FIP were reviewed for history, clinical signs, physical findings, and diagnostic test results. Sex, neutering status, breed, age, and month of onset of disease were compared with the characteristics of the clinic population. Age and neutering status were significantly correlated with FIP-suspicion. Sex, breed and onset month were not associated with FIP. There were many more FIP-suspected cases in cats in young cats or male intact cats. Effusion was observed in 85.8% of the FIP-suspected cats. Increased serum amyloid A (SAA) and lymphopenia were common laboratory abnormalities in the FIP cases. Furthermore, 91.7% of the cats highly suspected of FIP had an albumin/globulin (A/G) ratio < 0.6, while 85.3% had an A/G ratio < 0.5. The mortality rate for FIP-suspected cats was 67%, and six submitted cases were confirmed by FIP-specific immunohistochemistry. Of the 30 cats treated with GS-441524 and/or GC376, 29 were clinically cured. The study highlights the diverse range of clinical manifestations by clinicians in diagnosing this potentially fatal disease. A/G ratio and SAA were of higher diagnostic value. GS-441524 and GC376 were efficient for the treatment of FIP-suspected cats.

Feline infectious peritonitis (FIP) is a worldwide disease of domestic and wild felids. The pathogen of FIP is feline infectious peritonitis virus (FIPV), which is mutated from feline coronavirus (FCoV) 1,2 . Although much research into FIP has been performed, FIP remains one of the most prevalent and fatal infectious diseases of cats.
Except for non-specific clinical signs, such as lassitude, inappetence, fluctuating fever and weight loss, the clinical presentation of FIP is complex and variable. Different clinical features may be a result of variations in the virus and in the nature of the individual host's immune response. FIP can be divided into three types-effusive, non-effusive and mixed-according to the different clinical features that emerge, but these types can change over time in any individual cat. Effusive FIP is mainly manifested as ascites and/or pleural effusion, while non-effusive FIP is mainly manifested as granuloma formation involving the central nervous system, eyes and abdominal organs, and it does not produce body cavity effusion 3 . When effusive and non-effusive signs occur together, it is mixed FIP 4 . To our knowledge, there are few studies about the clinical and laboratory features of FIP in China.
Ante-mortem diagnosis of FIP is a controversial and complicated issue. It is more difficult to diagnose FIP definitively in vivo in cats without effusion as clinical signs and laboratory features are relatively vague. Immunofluorescence staining of FCoV antigen in macrophages is highly specific, but it is difficult to detected due to the low numbers of macrophages in the effusion. At present, the gold standard for the diagnostic of FIP is immunohistochemistry (IHC) to identify FCoV antigens in diseased tissues. However, this invasive method has low operability and requires professional detection equipment and personnel. Also, a negative IHC result does not exclude the diagnosis of FIP, adding to the challenges of diagnosing this disease 5 . FIP has an extremely high mortality rate and was once considered a terminal disease. In recent years, 3C-like protease inhibitor (GC376) and nucleoside analogue (GS-441524) have been proven to have a certain therapeutic effect on FIP [6][7][8][9] . The

Results
Risk factors. There was no significant correlation between FIP and sex (p = 0.083). However, FIP was significantly correlated with neutering status (p < 0.001), and male intact cats were more susceptible to disease. As for breed, whether classified as crossbred or purebred (p = 0.069) or specific to each breed (p = 0.246), no correlation with the FIP was found. There was a significant correlation between age and FIP (p < 0.001), and young cats were more susceptible. The age of the FIP-suspected cases ranged from 1 month to 8 years and 2 months, with an average age of 13.1 months and a median age of 8 months. Within these cases, 40.2% were under 6 months old, 67% were under 1 year old and 90.6% were under 2 years old; moreover, the prevalence of older cats was significantly lower than that of young cats ( Table 1). The correlation of onset month was studied for the first time, and the number of FIP-suspected cats and clinic population in each month was counted. The results showed that the highest prevalence was 1% in May. The prevalence in January (0.95%), February (0.87%), September (0.66%) and November (0.77%) were higher than the total prevalence, and the lowest prevalence was 0.35% in July. However, in the statistical analysis, there was no correlation between FIP and month (p = 0.135; Table 1).
Stressful events before the diagnosis were documented for 60/127 cats. Since some cats had multiple stressors, a total of 77 stressors were recorded. Of these events, changing environment was the most common factor leading to stress (40.3%), followed by new pets in the household, coexisting with other diseases, being frightened, changing food (Fig. 1). In addition, information on the housing density of 83 cats was collected, of which 43 (51.8%) were single-cat households and 40 (48.2%) were multi-cat households.
Clinical signs. Of the 127 cats highly suspected of FIP, 12 (9.4%) presented with non-effusive FIP and 109 (85.8%) presented with effusive FIP, of which 92 presented with ascites (84.4%), 11 with pleural effusion (10.1%), and 6 with both ascites and pleural effusion (5.5%). In addition, 6 cats (4.7%) presented with mixed FIP. Some presented with non-effusive FIP first, followed by ascites and the development of effusive FIP. Some cats showed effusive FIP first and develop neurological signs a few days before dying.
The clinical signs of cats highly suspected of FIP are shown in Table 2. The most common non-specific signs were weight loss (93.8%), lassitude (86.2%) and inappetence (86%). Icterus and fever were also present in more than half of the cases, 59.4% and 52.9%, respectively. Of the 85 cats highly suspected of FIP for which the temperature was documented on physical examination, 52.9% had a temperature above 39.5℃, 36 www.nature.com/scientificreports/ above 40℃ and 10.6% had a temperature above 40.5℃ (see Supplementary Fig. S1). In the results, 41.3% of the cats had a palpable mass on abdominal palpation, which could be either mesenteric lymph node enlargement or accretion of the intestinal wall. Diarrhoea was uncommon, occurring in only 17.6%, while dyspnoea occurred in 35.2%, and was most common in cats with pleural effusion. In addition, ocular and neurological signs generally occurred only in non-effusive or mixed forms of suspected FIP cases.
Haematology and serum biochemistry. As shown in Table 3 www.nature.com/scientificreports/ In serum biochemistry, the total protein (TP) levels were mostly normal (75.2%); a few (14.7%) cats had an increased level and a very few had a decreased level. More than half of the cats highly suspected of FIP showed hypoalbuminemia (58.7%). Although the albumin (ALB) levels of the remaining cats were normal, they were also near to the lower limit of the reference range. Hyperglobulinemia was also common in cats highly suspected of FIP (57.8%), and the remainder of the cats had normal globulin (GLOB) levels. The A/G ratio ranged from 0.137 to 0.767, with an mean of 0.399 and a median of 0.385. Among the results, 91.7% were lower than 0.6, 85.3% were lower than 0.5 and 54.1% were lower than 0.4 (see Supplementary Fig. S2). An increase in total bilirubin  www.nature.com/scientificreports/ (TBIL) was observed in nearly half of the cats highly suspected of FIP (48.7%), and alanine aminotransferase (ALT) and alkaline phosphatase (ALP) were mostly normal. Creatinine (CRE) decreased in 47.9% of the cats, while urea decreased in 60.7% of the cats and rarely increased. Amylase (AMYL) and lipase (LIPA) were mostly negative and only positive in fewer cats. In addition, feline serum amyloid A (SAA) increased in all cats (Table 3).
Imaging and molecular biology. Of the 127 cats, 90 were examined by ultrasound, which showed different degrees of ascites, and a few cats demonstrated loss of corticomedullary distinction within the kidneys. Radiography was performed in 21 cats, which showed pleural effusion and/or ascites (Fig. 2). Among 109 cats with effusive disease, 98 had extracted effusion. Then we detected FCoV RNA of 98 effusion by RT-PCR, 83 cats (84.7%) showing FCoV-positive. FCoV mutation was detected in 34 FCoV-positive effusions, with 12 successfully sequenced and partial S gene sequences obtained, all of which had M1058L or S1060A mutations. Furthermore, Rivalta's tests were performed on 85 cases with a positivity rate of 100%; FCoV antibody tests were carried out in 12 cases, and all were positive. Microscopic examination of smears of the effusions of 25 cats showed many pink protein granules and an increase in inflammatory cells.
Histopathology and immunohistochemistry. 13 cats underwent exploratory laparotomy, and histopathologic diagnosis was performed in 9 cats (see Supplementary Table S2 for details on 13 cats). After exploratory laparotomy, the samples were subjected to histopathological examination and we found the typical pathological changes of FIP, namely systemic vasculitis and pyogranulomatous lesions (Fig. 2). Tissue samples of six cats were randomly selected and sent to LABOKLIN GMBH & CO. KG for IHC. This testing revealed the tissue macrophages were positive-staining for the FCoV antigen, confirming the diagnosis of FIP (see Supplementary  Fig. S3-S8).
Treatment and outcomes. The outcomes of 88 cats were recorded on follow up. Among them, 59 cats eventually died, with a mortality rate of 67%. Thirty-two cats were euthanised directly. Twenty-six cats did not take any treatment or only received treatment for symptoms, and they eventually died naturally. Only one noneffusive FIP cat was treated with GS-441524 for 4 weeks. After remission of the disease, the owner chose to stop the drug, and the cat eventually relapsed and was euthanised. Moreover, 29 cats (25 effusive, 3 non-effusive and 1 mixed) were cured with GS-441524 (2-4 mg/kg, once a day for at least 4 weeks) or GC-376 (6-8 mg/kg, once a day for at least 4 weeks), with a cure rate of 33%. The most obvious signs were rapid reduction of ascites and/or pleural effusion (within 1 week), recovery of mental and appetite and weight gain. The longest survivor was an effusive FIP cat, which was diagnosed on 1 January, 2019 and has survived for more than a year.

Discussion
In the past, FIP was considered a fatal disease by researchers. Meanwhile, it is difficult to apply the IHC gold standard in the diagnosis of FIP in China. With the development of research, there were gradually found breakthroughs in treatment on FIP, such as studies on GS-441524 and GC376 [6][7][8][9]16 . But the side effects of these two drugs are not clear, and cannot be used for clinical treatment immediately. Therefore, the establishment of a more accurate diagnostic system is crucial for FIP in China. This paper reviewed 127 FIP cases in Wuhan, China, and studied the diagnostic methods of several hospitals. It emphasizes the clinical manifestations of clinicians in the diagnosis of the disease, and perfects the FIP diagnosis system. Data on 127 cases highly suspected of FIP were collected in this study, where 13 cats underwent exploratory laparotomy, 9 samples from the cats highly suspected of FIP underwent histopathology, and 6 samples underwent IHC, all of which were finally confirmed as FIP. Therefore, the inclusion criteria for the cases in this study were reasonable. There were also many investigations and studies on FIP based on the analysis of highly suspected cases, which could not be diagnosed by gold standard methods (IHC), but still provided a great reference value for the diagnosis and treatment of FIP [17][18][19][20][21][22] . In addition, the implementation of invasive diagnostic methods was difficult in clinical practice. Even cat owners who chose euthanasia could not accept autopsy after death. Therefore, when IHC is not available, how to diagnose a FIP-suspected case more accurately is a problem that veterinarians pay close attention to at present. www.nature.com/scientificreports/ In this study, the prevalence of cats highly suspected of FIP was 0.61%, and Riemer's research in Germany showed that the prevalence of FIP was 1.42% 14 . Earlier studies also showed that the prevalence of FIP was 0.02% in single-cat or double-cat households, while the prevalence was 5%-10% in catteries 23,24 . There was no correlation between sex and FIP. However, some studies indicated that male cats were more prone to FIP 25 . The prevalence of intact males was significantly higher than that of other groups, and the prevalence of neutered cats was less, consistent with Rohrbach's findings 26 . This may be because sex hormones, especially androgens, have a negative effect on the immune system, increasing the risk of virus proliferation and mutation 27 . As reported previously, this study observed that FIP was significantly associated with age, such that young cats were more susceptible, possibly because the immature immune system and various stressors often resulted in high viral loads in young cats. It is possible that the increased rate of unneutered cats may also have been due in part to the young age of infected cats, meaning that the disease occurred before the age of neutering. Purebred cats were previously considered to be more prone to FIP. However, a growing number of studies, including this study, have shown that the proportion of purebred cats was not excessive when compared with the clinic population 14 . This study also explored the regularity of the month of FIP onset, which was not considered in other FIP-related epidemiological surveys, and finally found no significant difference in the prevalence of FIP in each month of the year. This study recorded stressors in 60 cats, the most common of which was changing environment. There was no doubt that stress could inhibit immune function and make viruses more susceptible to mutation and proliferation 10 . Housing density was considered one of the major risk factors for FIP as well, and overcrowding could lead to virus mutation and disease development 14 . However, more of the cats highly suspected of FIP in this study came from single-cat households. This may be because these cats were exposed to FCoV whilst they lived in a multicat household during their infancy, and then moved into single-cat household prior to developing FIP.
Most cases of FIP presented as effusive (up to 85.8%), in which ascites accounted for the majority. FIP could affect systemic organs, most commonly the abdominal organs, including the intestine, mesenteric lymph nodes, liver, kidney, and spleen. Neurological and ocular signs were typical signs of non-effusive FIP. A retrospective study of 286 cats with neurological disease showed that FIP was one of the main causes of neurological disease in cats 28 . In this study, 52.9% of the infected cats had a temperature over 39.5℃, and 36.5% had a temperature over 40℃. Other research has shown that the proportion of cats with fever was larger, where 82% of the cats had a temperature of over 39.5℃ and 39% had a temperature of over 40℃; fever was also found to be more common in effusive FIP 14 . There were no cases of pericardial effusion, scrotal swelling or skin papules, which were not common but could not be ignored in routine examination. The clinical signs of FIP always change over time (such as the gradual production of effusion and changes in the fundus), so repeated clinical examination was highly important to avoid delay in diagnosis and treatment.
Changes in haematology and serum biochemistry could only increase or decrease the suspicion of FIP, but could not be used as a definitive diagnostic method. In this study, there was no obvious indication of hematological changes in the diagnosis of the FIP cases, with HCT decreased in 40.2% of cats, and LYM decreased in 50% of cats. However, many studies have shown that lymphocytopenia is the most common haematological abnormality of FIP induced by natural infection or experiment 29,30 , caused by virus-induced apoptosis of T-cells 31 . In this study, we guess the protein level testing plays an important role in diagnosing FIP cases. ALB decreased in 58.7% of the cats and GLOB increased in 57.8% of the cats, so when both occurred at the same time, TP often remained within the normal range. This was why only 14.7% of the cats showed an increase in TP. Both hyperglobulinemia and hypoalbuminemia could increase the suspicion of FIP, but the most important serum biochemical abnormality was the A/G ratio. The A/G ratio reduction occurred in almost all cases of FIP. Different studies had different results on the effective critical value of the A/G ratio 32 . According to the A/G ratios in this study (91.7% < 0.6, 85.3% < 0.5, and 54.1% < 0.4), the critical value could be considered as 0.5. When the A/G ratio > 0.8, the possibility of FIP was very small; when 0.5 < A/G ratio < 0.8, FIP remained a possibility; when A/G ratio < 0.5, FIP was highly suspected. In addition, FIP was accompanied by inflammatory response, resulting in increased protein concentration in the acute phase protein 33 . Therefore, we guess that the increase in SAA may assist in the diagnosis of FIP.
Compared with serum detection, the detection and analysis of body cavity effusion in cats has better predictive value 32 . The positivity rate of Rivalta's test was 100% in this study. A study indicated that the sensitivity of Rivalta's test was 91.3%, the specificity was 65.5%, the positive predictive value was 58.4% and the negative predictive value was 93.4% 34 . Rivalta's test is a cheap and rapid detection method that does not need expensive instruments and is simple to operate. In view of the good sensitivity and negative predictive value, Rivalta's test should be included in the diagnosis of every cat with effusion 15 . The positive rate of FCoV RNA detected by RT-PCR was 84.7%. However, all coronaviruses, including FCoV, often mutate and recombine, so RT-PCR designed for specific sequences could not amplify all FCoVs 35 , and the positive results only represented that the cat carried FCoV, not the exact FIPV.
In this study a total of 24 cats were treated with GS-441524, of which 23 were cured. Only 1 cat relapsed after 4 weeks of treatment and was eventually euthanised. Treatment with GS-441524 for more than 8 weeks was highly effective. Treatment with GC376 alone or in combination with GS-441524 was also effective. In fact, most of the final deaths were due to no treatment, symptomatic treatment alone or euthanasia. In addition, Pruijssers and Denison indicated that a combination of potent, broad-spectrum anti-coronavirus drugs can enhance efficacy and reduce the emergence of drug resistance 36 .
In conclusion, FIP was not correlated with sex, breed or month, but it was significantly correlated with age and neutering status. FIP occurred significantly more often in young cats or male intact cats. Effusive FIP was the most common clinical feature. A decreased A/G ratio, increased SAA and lymphopenia were common laboratory abnormalities. GS-441524 and GC376 were efficient for the treatment of FIP-suspected cats.