Introduction

Protein-losing enteropathy (PLE) describes a syndrome of endogenous protein loss from either the lymphatic or microvascular circulations, or the mucosal layer of the small intestine into the intestinal lumen, leading to abnormally low serum albumin concentrations. This is often accompanied by concomitant decreases in serum globulins and cholesterol. There are several recognized causes of PLE, including infectious agents, immune-mediated disease, and inflammatory and neoplastic processes, although immune-mediated disease in the form of inflammatory bowel disease (IBD) is considered the most common underlying etiology.¹ Intestinal lymphangiectasia, alimentary tract lymphoma, and histoplasmosis infection are the other most common causes.² In the scenario of IBD-induced PLE, it is thought that abnormal infiltrates of immune cells into the small intestinal mucosal and submucosal layers cause mucosal exudation leading to protein leakage and loss.¹ Overall, PLE carries a relatively poor prognosis. A recent meta-analysis of 23 studies with follow-up information on 445 cases showed that PLE caused disease-associated death in 54.2% of dogs.¹ Another PLE study retrospectively analyzed response to treatment with glucocorticoids either alone or in combination with a second-line immunosuppressive and showed median survival times of 85 days and 166 days, respectively, which were not significantly different, concluding that glucocorticoid treatment alone is appropriate for treatment of PLE. However, this study did not investigate the effect of diet on the progression or outcome of PLE.³

Lymphangiectasia is defined by the abnormal dilatation and occasional rupturing of the small intestinal terminal lymphatics, resulting in measurable and potentially severe protein and cholesterol losses. Lymphangiectasia is often seen in conjunction with PLE.¹ Lymphangiectasia can be primary in origin, and is often considered congenital, although clinical signs are not typically present at birth.⁴ Secondary lymphangiectasia caused by IBD or other etiologies, such as severe pancreatitis, small intestinal lymphoma, or fungal enteritis, is presumed to arise from lymphatic blockade leading to pressure overload and subsequent leakage of protein-and cholesterol-rich lymph.¹ Lymphangiectasia can also be seen secondary to portal hypertension, constrictive pericarditis, and right-sided congestive heart failure. Primary lymphangiectasia is traditionally treated with dietary modifications and anti-inflammatory steroids, whereas secondary lymphangiectasia is treated by addressing the underlying disease. ⁴ The indication for corticosteroids is supported by the fact that small intestinal biopsies from dogs with lymphangiectasia contain increased numbers of inflammatory cells in the lamina propria.⁵ However, it is not clear whether these inflammatory cells represent a primary immune-mediated condition or a secondary inflammatory process, and there are to date no immunologic studies in dogs showing that primary lymphangiectasia is an immune-mediated condition. The potential adverse effects of corticosteroid treatment, in particular muscle protein catabolism, hypercoagulability, and hyperlipidemia, could accentuate the clinical challenges of dogs with PLE. For example, it has been shown that dogs with PLE are already hypercoagulable before starting treatment with prednisone, likely due to the combination of loss of antithrombin combined with the increased coagulability associated with chronic inflammation.²⁰ A therapeutic approach that does not include corticosteroid treatment would be potentially advantageous.

Recently, it has been shown that a low-fat diet is helpful in cases of PLE with evidence of lymphangiectasia that have been unresponsive to prednisone treatment or that relapsed when prednisone treatment was reduced.⁶ The lacteal ducts of the intestinal villi provide an absorptive mechanism for dietary fats and proteins, and it is thought that by reducing the concentration of fats in the diet, the amount of circulating lipid in lymphatic fluid is reduced, which decreases lymphatic flow and lacteal pressure. Resulting engorgement of lacteals is reduced, leading to reduced protein loss.^7,17 Yorkshire terriers are considered to be a breed that is predisposed to lymphangiectasia,⁵ and a study showed that they can respond well to dietary therapy and might not require immunosuppression to achieve clinical remission. ⁷ A recent retrospective study evaluated associations between clinical characteristics of dogs with PLE and response to an ultra-low-fat home-cooked diet, concluding that dogs with higher clinical activity scores were less likely to respond to dietary monotherapy.¹⁹ However, owing to its retrospective nature, the clinical activity scores of the dogs had to be inferred from the medical record.

The goals of the current study were to determine if dogs of breeds other than Yorkshire terrier with presumptive PLE and ultrasonographic evidence of lymphangiectasia would respond to treatment with various low-fat diets as monotherapy, based on clinical activity score, and what clinical variables might be associated with outcome. A second goal was to evaluate if resolution of ultrasonographic evidence of lymphangiectasia would occur in dogs who achieved remission. Our hypotheses were that dogs of various breeds would respond in a similar manner to dietary monotherapy and that ultrasonographic resolution of lymphangiectasia would correlate with clinical remission of presumptive PLE.

Materials and Methods

This study comprised privately owned dogs with no known history of gastrointestinal disease, hypoalbuminemia (serum albumin <2.7 g/dL) as defined by the IDEXX Reference Laboratory at the time of study enrollment (May 2017 to November 2018), clinical signs of gastrointestinal disease (vomiting, weight loss, decreased appetite, or diarrhea), and ultrasonographic evidence of presumptive lymphangiectasia (small intestinal hyperechoic linear striations).¹⁰ Owners were offered enrollment in a study of 6 mo duration at the discretion of the attending clinician. The minimum diagnostic database required for enrollment included a complete blood count (CBC), serum biochemistry, and full abdominal ultrasound by a board-certified radiologist. Dogs with clinically relevant abnormalities detected on hematology, biochemistry, or abdominal ultrasonography that were not attributable to PLE, or with history of a concurrent disease known to be associated with hypoproteinemia such as neoplasia, protein losing nephropathy, malnutrition, or end-stage liver disease, were excluded. Dogs who had been treated with prednisone or other immunomodulatory agents within the previous 30 days were also excluded. Informed consent was obtained from all owners of dogs who elected to participate. All dogs were medically managed by their owners in the home environment.

All cases were grouped upon study completion or study exit according to outcome—outcome group LOF: clinical remission on low-fat diet alone; outcome group LOP: clinical remission on low-fat diet with prednisone; outcome group TXF: failure to improve. Clinical remission was defined as a Canine Chronic Enteropathy Clinical Activity Index (CCECAI, previously described)⁸ score of 3 or less. Briefly, the CCECAI uses a questionnaire format combined with clinical variables to evaluate a dog’s gastrointestinal status based on a scoring rubric of 0–3, where 0 is normal and 3 is most abnormal. Appetite, activity level, stool quality, stool frequency, occurrence of vomiting, albumin level, weight, presence of ascites, and pruritus are all assessed. The sum of scores represents the CCECAI, which can be used to categorize levels of disease as insignificant, 0–3; mild, 4–5; moderate, 6–8; severe, 9–11; and very severe, >12.⁸

Dogs were initially enrolled into group LOF on an exclusive low-fat diet. Dogs were evaluated at five time points: enrollment (time point 0), 2 wk, 4 wk, 13 wk, and 26 wk after enrollment. At each evaluation, a CCECAI score was calculated with the participation of the owner, and a nonfasting serum biochemistry was collected. If clinical improvement was judged to be sufficient based on CCECAI score, blood work, and physical examination, pharmaceutical intervention was not recommended. If clinical improvement was not deemed adequate by owners or the attending clinician, then immunosuppressive oral prednisone was initiated at a dose approximating 2 mg/kg/day.

Ultrasound was performed at time points 0 and 26 wk by one of two board-certified radiologists (J.T.S. or A.T.W.). Ultrasound at time point 0 was a full abdominal study, which was also used to help rule out other causes for hypoproteinemia. Ultrasound at time point 26 wk was limited to examination of the gastrointestinal tract and screening for the presence of abdominal effusion. All ultrasounds were performed on dogs fasted at least 8 hr.

Owners were advised to choose between two commercial low-fat diets—either canned or kibble (Hill’s i/d Low-fat^a ; Royal Canin Gastrointestinal Low-fat^b)—or, if they preferred, one of two home-cooked low-fat diet options adapted from published veterinary recipes.⁹ The allowed commercial diets were reported to have between 18.1 and 23.0 grams of fat per 1000 kcal, according to their manufacturers. The recommended home-cooked diets ranged from 13.6 to 14.4 grams of fat per 1000 kcal. Owners were not discouraged from mixing ingredients from any of the dietary options offered (Table 1) to help with periods of hyporexia or anorexia, although the stated objective was to maintain all dogs on a consistent diet throughout the study. Once a diet was selected, recommended amount to be fed was calculated and provided to the owners based on the following standard equations: for active dogs, 130 kcal × BW_kg^0.75, and for inactive dogs, 95 kcal × BW_kg^0.75, where BW represented the dog’s approximate ideal body weight.

TABLE 1 Nutrient Profiles of Dietary Options

View Fullscreen

Results

Fourteen dogs met inclusion criteria. Breeds represented included Yorkshire terrier (n=3), mixed-breed dog (n=3), dachshund (n =2), Chinese crested, Cavalier King Charles spaniel, Portuguese water dog, Labrador retriever, Boston terrier, and Maltese. The median age was 9.5 yr (4–13 yr). Ten dogs were spayed females and four were castrated males. The median presenting CCECAI score was 11 (7–16), consistent with severe disease. All dogs presented with a history of abnormal stool consistency, decreased appetite, vomiting, or a combination of these signs. Twelve dogs presented with a history of abnormal stool consistency: seven with severe watery diarrhea, two with very soft stool, and two with slightly soft stool. Ten dogs presented with a history of decreased appetite: two severely decreased, six moderately decreased, and two slightly decreased. Six dogs presented with a history of vomiting: one severe (>3×/wk), two moderate (2–3×/wk), and three mild (1×/wk).

Initial clinicopathologic evaluation included a nonfasting serum biochemistry in all dogs and a CBC in 13 of 14 dogs (Table 2). For 1 dog, the CBC was not completed because of an oversight by the clinician in following study protocol. Abnormal biochemistry results included a median albumin of 1.2 (0.8–2.0) g/dL (reference range 2.7–3.9), median globulin of 2.0 (1.4–2.9) g/dL (reference range 2.4–4.0), median cholesterol of 105 (32–143) mg/dL (reference range 131–345), and median total calcium of 6.35 (3.7–8.5) mg/dL (reference range 8.4–11.8). All other median biochemistry results were within normal limits. Abnormal CBC results included a median neutrophil count of 14.2 (1.6–169.6) thou/μL (reference range 2.94–12.67), and a median platelet count of 530 (232–801) thou/μL (reference range 143–448). All other median CBC results were within normal limits. No dogs had azotemia based on blood urea nitrogen and creatinine, or elevated liver enzymes based on alanine aminotransferase and alkaline phosphatase, except for 1 dog with a mildly elevated alkaline phosphatase at 194 U/L (reference range 5–160).

TABLE 2 Presenting (Time Point 0) Continuous and Discrete Variables Across Outcome Groups

View Fullscreen

Based on clinician discretion, eight dogs received courses of additional gastrointestinal medications including metronidazole, cyanocobalamin, probiotics, famotidine, maropitant, and natural clay, not including rescue therapy for TXF dogs. Duration of therapy with metronidazole ranged from 6–10 days (n=3) to a course of 10 days followed by a course of 14 days (n=1). Dogs started on cyanocobalamin continued to receive it in injectable or oral form throughout the study (n=2). Courses of probiotics ranged from 10 days (n=1) to 28 days (n=1). One dog received previously prescribed famotidine throughout the study. One dog received an 8-day course of maropitant. Courses of natural clay ranged from ~2 mo (n=1) to ongoing throughout the study (n=1).

Six dogs (43%) achieved clinical remission on dietary monotherapy and were assigned to group LOF. Eight dogs (57%) were deemed to have an insufficient response to dietary monotherapy and were started on immunosuppressive prednisone. Five dogs (36%) on combined prednisone and low-fat dietary therapy achieved clinical remission and were assigned to group LOP. Overall, 11 of 14 dogs (79%) went into clinical remission. Three dogs failed therapy despite addition of prednisone and subsequent rescue protocols and were assigned to group TXF. The three dogs that failed therapy were euthanized before study completion. One LOP dog did not return for its final recheck examination because of owner noncompliance, but data up to censorship were included in this report based on the owner’s report via telephone that the dog remained in clinical remission with no ongoing gastrointestinal signs.

The decision to initiate prednisone therapy for the eight dogs that were deemed to be failing dietary monotherapy was based on failure to improve clinically, as reflected in the CCECAI score. In all cases in which the CCECAI score was found to be increased at one of the regular recheck appointments, despite compliance with the study protocol, the addition of immunosuppressive steroid therapy was recommended. The common pattern that emerged in almost all of these dogs (seven of eight) was initial improvement on dietary monotherapy followed by a relapse of clinical signs with a corresponding increase in CCECAI score. The one dog that did not fit this pattern, a 5 yr old female spayed Cavalier King Charles spaniel, was taken to her primary care veterinarian only 7 days after study initiation for diarrhea, at which time the owner requested that her pet be placed on a steroid. Steroid therapy was initiated, and this dog went on to achieve clinical remission and was successfully weaned off steroids during the last 6 wk of the study. A more typical example, a 5 yr old female spayed cocker spaniel mix, achieved marked clinical improvement by wk 4, with a CCECAI score of 0, down from 10, and an albumin of 2.2 g/dL, increased from 1.0 g/dL. This dog relapsed during the following weeks and at the 13 wk recheck had a CCECAI score of 14 and albumin of 1.1 g/dL. Immunosuppressive prednisone was started, but this pet ultimately failed therapy and was euthanized for poor quality of life despite an aggressive rescue protocol. In most cases, prednisone therapy was initiated at the 4 wk recheck (n=3) or at the 13 wk recheck appointment (n=4). All surviving dogs that were started on prednisone, other than the one dog that was weaned as described above, continued to receive a tapering dose of prednisone as of the time of study completion.

LOF dogs included one of each of the following breeds: Yorkshire terrier, Chinese crested, Portuguese water dog, Labrador retriever, Boston terrier, and a German shepherd mix. Dogs that failed therapy included a dachshund, a Maltese, and a cocker spaniel mix. All Yorkshire terriers (n=3) went into remission, with two (67%) in outcome group LOP and one (33%) in group LOF. However, there were no significant differences in outcomes based on breed.

Eight owners elected to feed one of the low-fat home-cooked diets: seven of these dogs achieved clinical remission, with four in outcome group LOP, three in outcome group LOF, and one in outcome group TXF. Four elected to feed a low-fat diet from Hill’s Nutrition^a (Hill’s Prescription Diet i/d Low Fat Dry Dog Food or Hill’s Prescription Diet i/d Low Fat Wet Dog Food); two of these dogs achieved clinical remission, with one in outcome group LOP, one in outcome group LOF, and two in outcome group TXF. Two elected to feed a low-fat diet from Royal Canin^b (Royal Canin Canine Gastrointestinal Low Fat Dry Dog Food or Royal Canin Canine Gastrointestinal Low Fat Canned Dog Food): both dogs achieved clinical remission, with one in outcome group LOP and one in outcome group LOF. There were no significant differences in outcome group based on diet selected (P=.085). There were likewise no significant differences in albumin levels (P=.34) or CCECAI scores (P=.45) based on diet selected.

Owners were offered gastroduodenal endoscopy and biopsy for further characterization of their dogs’ disease, but only two dogs had endoscopic biopsies collected upon study enrollment. Two of the three euthanized dogs had postmortem gastrointestinal biopsies collected. Of the two dogs that had endoscopy performed on study entry, one had histologic evidence of cryptal dilation and mild lacteal dilation without an abnormal lymphoplasmacytic infiltrate. This dog improved on dietary monotherapy. The second dog had histologic evidence of both lymphoplasmacytic inflammation and severe lacteal dilation. This dog improved on dietary therapy with prednisone. Both dogs in group TXF (euthanized) who underwent postmortem gastrointestinal biopsy had histologic evidence of widespread, moderate to marked dilatation/ectasia of small intestinal glands and crypts.

There were no significant differences in the continuous/discrete variables age, weight, or initial CCECAI score on presentation between dogs in outcome groups. Significant differences in clinicopathologic variables between outcome groups at presentation included globulin levels, which were higher on presentation in LOF dogs compared with LOP dogs (P=.03), and total calcium, which was also higher in LOF dogs compared with LOP dogs (P=.003). There were no other significant associations between wk 0 clinical variables and outcome (Table 2).

Ten of the 11 dogs that went into clinical remission had ultrasounds of the gastrointestinal tract performed at the 26 wk recheck examination. Of these 10 dogs, 4 had resolution of linear striations and 6 had ongoing evidence of this finding (Figure 1). Dogs with resolution of linear striations had a lower albumin/globulin ratio (0.6 versus 0.8, P=.02) on initial presentation than dogs without resolution. No other clinical variables were significantly associated with resolution of linear striations.

View Figure

• Download as Powerpoint
• Download Figure

LOF dogs achieved a significant reduction in CCECAI scores within 2 wk of starting treatment (P < .0001), including 5 out of 6 dogs achieving a CCECAI score of 3 or less. LOP dogs initially responded to diet monotherapy, with a reduction in CCECAI score by wk 4 (P=.0008), but then relapsed clinically. With addition of prednisone therapy, LOP dogs did ultimately achieve clinical remission. TXF dogs appeared to have initial improvements in CCECAI score, with a mean score of 5.33 at wk 4, which was not statistically significant. CCECAI scores for TXF dogs then worsened as they failed treatment (Figure 2). At wk 13, CCECAI scores were higher for TXF dogs compared with LOF and LOP dogs: 2.67 ± 0.99 for LOF dogs (P=.001) versus 3.80 ± 2.3 for LOP dogs (P=.004) and 11.5 ± 2.5 for TXF dogs.

View Figure

• Download as Powerpoint
• Download Figure

Albumin levels were not significantly different between groups at presentation. Albumin increased compared with time 0 for LOF dogs at the 2 wk recheck (P=.01), 4 wk recheck (P=.006), and 13 wk recheck (P=.008). Albumin in LOP dogs compared with time 0 improved at the 13 wk recheck (P=.01) and was also improved at the 26 wk recheck (P < .0001) (Table 3).

TABLE 3 Clinicopathologic Variables by Outcome Group Across All Time Points

View Fullscreen

At wk 2, the mean albumin for LOF dogs was higher than the mean albumin for LOP (P=.008) and TXF (P=.008) dogs. As would be expected, at wk 13 the mean albumin for LOF dogs (P=.009) and LOP dogs (P=.03) was higher than the mean albumin for TXF dogs, the nonresponders (Figure 3).

View Figure

• Download as Powerpoint
• Download Figure

Globulin levels increased across all groups at the 4 wk and 13 wk rechecks (P=.03). Cholesterol levels increased in LOP dogs at the 4 wk (P=.05) and 26 wk (P < .001) rechecks. Calcium levels increased in LOP dogs at the 4 wk (P=.002) and 13 wk (P=.01) rechecks (Table 3).

Discussion

Dietary therapy for PLE and specifically for lymphangiectasi-aassociated PLE has been a long-standing clinical recommendation.⁵ However, to the authors’ knowledge, using fat-restricted diet as monotherapy has only been investigated thus far retrospectively. Our results suggest that dietary monotherapy as treatment of lymphangiectasia-associated PLE can be successful in various breeds.

In the current study, clinical remission was defined as a CCECAI score of 3 or less. The median presenting CCECAI score was 11 (severe disease), and 11 of 14 dogs (79%) were in clinical remission at the end of the 6mo study period. Of these 11 dogs, 6 were treated successfully with low-fat diet monotherapy, and 5 were treated initially with low-fat diet with subsequent additional treatment with prednisone when clinical improvement was considered inadequate on diet alone (Figure 2). One possible explanation for the failure of some dogs to improve on dietary monotherapy is that their suspected lymphangiectasia was a less significant cause for their PLE in comparison to presumptive idiopathic or immune-mediated inflammatory lesions of the intestinal mucosa (IBD). In this scenario, lymphangiectasia would be considered a secondary pathology to disruption of normal function by inflammatory mediators,¹ and a better result might have been achieved by provision of a novel protein or hydrolyzed diet that was also low in fat. It is also possible that for certain dogs, the low-fat diet option selected was still not low enough in fat to attenuate their disease.

Hyperechoic small intestinal mucosal striations on abdominal ultrasound have been shown to be associated with histologic lacteal dilation in 96% of dogs and clinically with PLE in 76% of dogs.¹⁰ Gaschen et al. reported that hyperechoic mucosal striations in the jejunum had a sensitivity of 75% and a specificity of 96% for dogs diagnosed with PLE based on clinical and histological evidence. They further showed that 100% of dogs with PLE showed abnormal echogenicity in the duodenal mucosa on presentation, compared with 58.3% of dogs with IBD and 34.5% of dogs with food-responsive diarrhea.¹¹ These results in conjunction with our findings support the clinical approach that for dogs with clinicopathologic evidence of hypoproteinemia, and for whom other causes of protein loss have been excluded (although exclusion of other causes of protein loss was incompletely performed in the current study), abdominal ultrasound is a reasonable diagnostic modality for identifying dogs with dilated lacteals and presumptive PLE. Although histologic evidence of lymphangiectasia is the gold standard diagnostic criteria, many owners are not willing or financially able to pursue endoscopic biopsies, and a subset of dogs with comorbidities such as respiratory or heart disease could also not be good candidates for the required anesthesia.

Four of the 10 dogs that went into clinical remission also showed evidence of resolution of mucosal linear striations on 6 mo follow-up ultrasonographic examination (Figure 1). The only clinical variable significantly associated with ultrasonographic resolution was a decreased albumin/globulin ratio, which is of unknown significance and could represent a type 2 statistical error. Three of the 4 dogs who achieved ultrasonographic resolution were in the LOF group. It can be theorized that the remaining 6 dogs that achieved remission, while still ultrasonographically abnormal, could have improved at a subtler histopathologic level. For example, histologic changes in dogs with diet-responsive IBD are often not seen,⁴ or have been shown to be present but quite subtle and difficult to differentiate even on histopathology, evidenced by ultrastructural changes such as improved microvillus height and mild reductions in the density of inflammatory cells in the lamina propria.¹² Histologic evaluation would be required both before and after dietary therapy for lymphangiectasia-associated PLE to investigate whether similarly subtle distinctions might be present.

Hypoalbuminemia is the hallmark clinicopathologic abnormality in dogs with PLE and is associated with significant morbidity.⁵ LOF dogs showed significant improvement in serum albumin at their 2, 4, and 13 wk rechecks, but serum albumin at wk 26 remained static at a median of 2.1 (1.6–2.7) g/dL. In contrast LOP dogs achieved a significant improvement in albumin at wk 13 and were also significantly improved at wk 26 with a median of 2.6 (2.3–2.9) g/dL (Figure 3). These results could suggest that dogs on dietary monotherapy reached a plateau in recovery, as their hypoalbuminemia was no longer improving at the 26 wk recheck despite significant improvements at all other time points. A longer follow-up period and a larger cohort of test subjects would be required to investigate this suspicion.

Curiously, one of the LOF dogs that achieved clinical remission with a final CCECAI score of 3 did not have concurrent normalization of albumin levels. This dog, an 11 yr old female spayed mixed-breed dog weighing 21.0 kg, had an albumin of 1.2 g/dL at time point 0, which only increased to 1.6 g/dL at 26 wk. Nevertheless, the owner reported dramatic improvements in the dog’s appetite, energy level, stool quality, and overall quality of life. This case points out that clinical outcome is not always correlated with albumin levels and that escalation of treatment may not be required to address blood test results when the patient is otherwise prospering.

Hypocholesterolemia is common in PLE, suggesting malabsorption and/or the leakage of lipoprotein-rich lymph, and the presence of hypocholesterolemia in dogs with hypoalbuminemia helps to differentiate PLE from protein-losing nephropathy, which would typically present with hypercholesterolemia.⁴ It has been shown that worsened hypocholesterolemia 1 mo after initiation of immunosuppression is associated with shorter survival in dogs with PLE.¹⁴ An increase in cholesterol concentration was significantly associated with a reduction in the hazard of death in dogs with PLE in a univariable Cox proportional hazards regression model, but these results were not confirmed in a multivariable model; thus, the effect of serum cholesterol concentration on survival in dogs with PLE was not definitively confirmed.¹³ To the authors’ knowledge, no studies have thus far shown an association between improved cholesterol levels and resolution of PLE. The median presenting cholesterol in this study was 105 (32–143) mg/dL, and 11 of 14 dogs (79%) were hypocholesterolemic on presentation. The three dogs with normal presenting cholesterol levels were at the low end of the normal range (139, 143, and 143 mg/dL). Cholesterol levels only increased in LOP dogs; however, this result must be interpreted in light of the fact that a known prednisone side effect is increased plasma cholesterol. There were no significant changes in cholesterol levels for LOF or TXF dogs.

Hypoglobulinemia is also commonly identified in dogs with PLE,⁴ although many dogs with PLE have normal or increased globulins due to comorbidities causing increased globulin production.⁵ The median serum globulin for dogs on presentation was decreased at 2.0 (1.4–2.9) mg/dL, and globulin levels increased across all outcome groups at the 4 wk and 13 wk rechecks (P=.03), suggesting that both dietary monotherapy and low-fat diet combined with prednisone therapy could help restore serum globulin levels. LOF dogs that went into remission on exclusive dietary therapy had higher median serum globulins of 2.5 (1.8–2.9) versus 1.7 (1.5–2.1) mg/dL (P=.03) on presentation than LOP dogs. These results suggest that more severely clinicopathologically affected dogs could be less likely to respond to the low-fat diet options offered in this study (Table 1). This was also the conclusion of the investigators who retrospectively evaluated predictors of outcome in dogs with PLE treated with ultra-low-fat diet, based on their finding of lower initial clinical activity scores in the dogs that responded to dietary monotherapy.¹⁹ This information is intriguing, and future studies with higher case numbers might serve to guide clinicians in their choice of therapy for lymphangiectasia-associated PLE based on clinicopathologic severity.

It is unknown why the three TXF dogs in this study failed both dietary and immunosuppressive therapy for presumptive PLE despite evidence of presumptive lymphangiectasia on abdominal ultrasound. It is curious that these dogs showed initial clinical improvement with a median CCECAI score of 5 on wk 4, before worsening (Figure 2). Rescue protocols used invariably included immunosuppressive prednisone (~2 mg/kg/day) and variably included dexamethasone, metronidazole, ondansetron, omeprazole, maropitant, or capromorelin depending on the specifics of the individual case. One dog received a second immunosuppressive (cyclosporine). No significant association was found between the TXF outcome and any of the clinical variables collected at presentation, including CCECAI score, albumin level, or presence of linear striations on ultrasound. Thus, no predisposing factor that might predict treatment failure can be implicated from the results of this study.

For two of the three TXF dogs, postmortem gastrointestinal biopsy was performed. In the first dog, a 7 yr old female spayed cocker spaniel mix, there was widespread, moderate to marked dilatation/ectasia of small intestinal glands and crypts seen, while neither lymphangiectasia nor neoplasia was detected and lymphoplasmacytic inflammation was mild. This dog had initially improved on dietary monotherapy, achieving a perfect CCECAI score of 0 at the 4wk recheck, but then relapsed severely and had a CCECAI score of 14 at the 13 wk recheck. This unfortunate outcome suggests that although dietary therapy was initially effective in resolving this dog’s clinical signs, more deep-seated disease eventually led to relapse. The second dog, a 6 yr old male castrated Maltese, also had moderate to marked cryptal dilation identified with mild to moderate lymphoplasmacytic enteritis and mild to moderate lacteal dilation. These results prompted the pathologist to remark that the small intestinal lesions were highly similar to those of the first dog, and he speculated that “given the high histologic and clinical similarities between the present case and the previous report, it is considered possible that these histologic features may be associated with a condition that is more refractory to currently used medical management protocols” (Z. Demeter, Pathology report for 2 site biopsy, IDEXX Reference Laboratories 2019, Accession number 2001293713). The second dog also showed some initial improvement clinically with a reduction in his CCECAI score from 14 to 8 at the 2 wk recheck but then relapsed back to a CCECAI score of 11 at the 4 wk recheck. A retrospective study of 58 dogs with idiopathic chronic enteropathies found that dogs with crypt abscesses were likely to have more severe intestinal protein loss, more severe clinical disease, and shorter survival times.¹⁶ It is thought that crypt dilation, if severe enough, could apply a compressive force to adjacent villi, resulting in lymphatic obstruction.¹⁸ The clinical relevance of crypt lesions in dogs with PLE merits further investigation.

No association was found between the low-fat diet selected and outcome variables including outcome group, albumin level, and CCECAI score. However, owners were allowed to choose between home-cooked and commercial low-fat diet options with different macronutrient profiles (Table 1), and, in the spirit of compassionate care, owners were not discouraged from mixing or blending ingredients from the home-cooked and commercial diet options available to help with periods of hyporexia or anorexia. These variances in the low-fat diets provided limit the applicability of the statistical analysis that looked for association between diet selected and outcome in this study. Strict compliance with a single dietary option would have provided more easily interpreted results. It is also possible that a completely different approach to diet might have been helpful to the nonresponders, e.g., more fat restricted, novel protein with fat restriction, or hydrolyzed, but this study did not investigate these dietary options.

Another recognized limitation of the study was the lack of participation by, or consultation with, a board-certified veterinary nutritionist. The home-cooked diets offered to the owners were not complete and balanced, and precise instructions on how to prepare each ingredient were not provided. Rather than rely on published recipes, that despite being formulated by a veterinarian were not Association of American Feed Control Officials compliant, a veterinary nutritionist could have reformulated these recipes to ensure that micronutrient levels met Association of American Feed Control Officials and National Research Council standards and to provide greater standardization on how ingredients should be prepared. Although owners were advised on how to further supplement the home-cooked diets to make them complete and balanced once the study was completed, it is possible that nutrient deficiencies and variations in food preparation of the home-cooked diets affected outcome in unforeseen ways.

As the study duration was only 6 mo, it is unknown whether the clinical remissions achieved remained durable. It is thought that many dogs who achieve remission from lymphangiectasia will eventually succumb to the disease, even with remissions that could last years.⁴ It would be of great value to collect longer-term outcome data on a larger cohort of dogs treated successfully with low-fat diet to determine the extent of their remission.

It could be argued that the clinical improvements seen with dietary modification in some dogs in the current study were not due to fat restriction but instead were a serendipitous side effect caused by elimination of one or more food antigens that were causing an adverse food reaction leading to inflammation and secondary lymphatic obstruction. However, this is considered unlikely as provision of a strict elimination diet, the usual approach to antigen removal in diet-responsive enteritis,⁸ was not a feature of the current study. Furthermore, dogs with food-responsive enteritis do not typically present with hypoalbuminemia and severe clinical signs like the dogs in this study.⁸ An exception is the familial syndrome of soft-coated wheaten terriers, where food allergy–induced enteritis can progress to PLE. Affected soft-coated wheaten terriers have been shown to experience significant decreases in serum albumin after a provocative food trial, although never falling below the reference range.¹⁵

The rate of remission of presumptive PLE in the current study (79%) was higher than is typically seen in PLE.¹ However, dogs were offered study enrollment at the discretion of the attending clinician; thus, it is possible that dogs with more severe protein-losing disease were less frequently considered for inclusion, leading to a better overall outcome in the study population than in a population of dogs selected randomly. Although this possible selection bias could not be controlled, the median presenting CCECAI score (11) and median presenting albumin levels (1.2 g/dL) argue that the included dogs nonetheless had severe disease. Dogs with PLE caused by lymphangiectasia have been shown to have a 22.9% increased hazard of death for each unit increase in CCECAI score, and those with a CCECAI score >8 have been shown to have a median survival time of only 109 days.¹³

This study has several additional limitations, including small sample size, which could have led to type 2 statistical errors. Serum chemistry results were based on nonfasting samples because maintaining the highest possible plane of nutrition was deemed more important. However, this could have decreased the accuracy of the results, particularly for serum cholesterol levels, which may have been artificially increased postprandially. This study did not follow a rigid treatment protocol, as different diets were consumed, and some dogs received short courses of additional medications that could have influenced outcome. Owing to financial constraints, preprandial and postprandial bile acid assays (to rule out hepatic functional insufficiency), urinalysis and urine protein:creatinine ratios (to rule out protein-losing nephropathy), and baseline cortisol measurement (to rule out Addison disease) were not routinely performed. The initial diagnosis of presumptive PLE instead relied on clinical signs, serum biochemistry, abdominal ultrasound, and clinician experience. As such, there could have been unappreciated protein loss due to undiagnosed disease processes. In retrospect, the authors regret that the inclusion criteria did not require these additional assays to rule out other possible causes of hypoalbuminemia. Although the study goals included making the diagnostic process as cost-effective as possible, these additional tests would have been a relatively inexpensive addition to the protocol and would have lent more credibility to the presumed diagnosis of PLE.

Also, and notably, small intestinal biopsies were not consistently collected in this study, instead depending on abdominal ultrasound for the presumptive diagnosis of lymphangiectasia, which is less accurate. Although a significant limitation, it is also a pillar of what this study intended to demonstrate—that a less expensive/invasive modality might be pursued to establish a presumptive diagnosis. The favorable clinical response to therapy achieved in a high percentage of dogs in this study supports using ultrasound to drive treatment decisions for lymphangiectasia-suspected PLE when gastrointestinal biopsy cannot be performed. Because of ethical reasons, this study did not include a control group and was therefore not blinded or randomized. Conclusions must therefore be interpreted cautiously, and further objective clinical trials are needed to confirm the findings.

Conclusion

Low-fat diet appears to be an effective monotherapy in some dogs with presumptive protein-losing enteropathy and presumptive lymphangiectasia, regardless of breed. In dogs that do not respond to low-fat-diet monotherapy, the addition of prednisone can also lead to remission, although the durability of remission with either an exclusive low-fat diet or a low-fat diet in conjunction with prednisone is unknown. Ultrasonographic evidence of dilated lacteals can be used to develop a treatment plan based on a presumptive diagnosis of lymphangiectasia; however, this carries the risk of misdiagnosis and undertreatment of dogs that have other causes of PLE that might only be found on biopsy.