SSc is a chronic multi-system disease with wide-reaching consequences. Gastrointestinal features are present in over 90% of cases and these, together with other disease manifestations, may lead to nutritional decline. This produces substantial morbidity, including reliance on enteral support and even parenteral nutrition-dependent intestinal failure. These complications carry an associated mortality. Up to 18% of patients with SSc are reported to be at high risk of malnutrition [as assessed by Malnutrition Universal Screening Tool (MUST) criteria], with risk increasing with disease severity. Little is known about this decline, its rate of progression and how it affects the individual. Few case series report on nutritional interventions. Most current interventions are based on experience in other diseases. The development of specialist knowledge of SSc-related gastrointestinal disease management and nutritional screening and interventions is required. This paper reviews current knowledge relating to malnutrition and its management in SSc.
SSc is a chronic multi-system disease with a reported prevalence of 8.2/100 000 in the UK . It is characterized by immune activation, which results in fibrosis, vascular abnormalities and neural dysfunction [2–4]. Although fibrosis characteristically affects the skin, other internal organs, including the gastrointestinal (GI) tract, may be affected.
Symptoms of varying severity, suggestive of GI tract involvement, are reported by more than 90% of patients with SSc . This, in addition to other influences, may result in malnutrition, which is variably described, depending on the assessment method. Clinical assessment reports medium to high malnutrition risk in 28% of patients, while bioelectrical impedance demonstrates malnutrition in 56% of patients [6, 7]. Malnutrition has a negative impact on quality of life, as well as producing substantial morbidity with, in the severest cases, the development of parenteral nutrition (PN)-dependent intestinal failure (2%) [6, 8]. Previously, malnutrition was estimated to be responsible for 12% of SSc deaths (1972–1977) . Recently, malnutrition-related deaths have fallen, possibly due to improved nutritional support, but still account for up to 4% of all deaths (1997–2001) .
Causes of nutritional impairment in SSc
Many factors may contribute to the development of nutritional impairment.
Depression and anxiety may increase or decrease appetite. Depression is variably reported to affect 19–69% of patients with SSc [10, 11]. Not only is this far higher than in the general adult UK population (2.6–3.2%), but it is also much higher than in individuals with two or more common chronic physical diseases (23%) [12, 13]. Anxiety is reported in 49–80% of patients with SSc, compared with 4.4% of the general adult UK population [10, 11, 14]. Depression and/or anxiety may be associated with nutritional impairment through reduced appetite or motivation for key nutritional tasks, including shopping and meal preparation .
SSc may affect functional status due to cutaneous and/or visceral disease manifestations. Digital ulcers occur in up to 58% of patients, causing pain, infection and irreversible tissue loss . Hand disease may have a negative impact on the performance of daily activities including food preparation and eating . Well-being, working ability and daily activities may also be influenced by other disease manifestations. Medsger lung severity scores, fatigue severity scores and disease-related myalgia, for example, are independently associated with global disability, thereby potentially impairing activities including meal preparation and consumption [18–20].
Oropharyngeal manifestations occur frequently in SSc, and may be associated with reduced oral intake [21, 22]. Microstomia and xerostomia have been reported in two-thirds of patients . The Canadian Scleroderma Research Group (CSRG) evaluated 586 patients and noted significant correlation between oral aperture size and malnutrition risk on bivariate, but not multivariate, analysis . Xerostomia predisposes to oral pathogens, which, in conjunction with poor hand dexterity, may hinder effective dental hygiene [23–25]. This increases the risk of dental caries and subsequent tooth loss, impairing mastication and altering food choice [25, 26]. It has yet to be shown whether this leads to malnutrition.
GI dysmotility is arguably the most significant problem in SSc but the pathogenesis remains poorly understood. Initially, progressive fibrosis of submucosal and muscle layers throughout the GI tract was thought to be the principle cause . More recent evidence suggests that autonomic neuropathy is an early component and may even precede fibrosis [4, 27].
The most commonly reported symptoms originate from the upper GI tract . Of 402 patients questioned [69% limited cutaneous (lcSSc), 30% diffuse (dSSc)], 94% reported upper GI symptoms, mostly refractory to medical therapy . Seventy-one per cent of patients were on acid suppressants, including proton pump inhibitors (PPI), but many still experienced reflux symptoms . Oesophageal manometry and endoscopy in 60 unselected patients (80% with upper GI symptoms) revealed the following: hypotensive lower oesophageal sphincter (95%), aperistalsis (41%), oesophagitis (60%) and Barrett’s oesophagus (18%) . Cross-sectional imaging often reveals distal oesophageal dilation, which correlates with dysmotility [29, 30]. Reduced peristalsis leads to dysphagia. The associated impaired clearance of refluxed acid can, in turn, lead to peptic stricturing. Prior to widespread PPI use, 40% of patients with SSc with erosive oesophagitis developed strictures . Endoscopic evidence of oesophagitis may still occur with PPIs . However, strictures are now infrequently reported but still need to be sought since they are amenable to dilatation [32, 33].
Symptomatic oesophageal dysmotility would be expected to cause nutritional impairment. Surprisingly, however, studies fail to show any significance. This may simply reflect their small patient numbers or short evaluation periods. A brief dietary study involving 30 patients (13 lcSSc, 17 dSSc) with GI symptoms (23 with oesophageal symptoms) and 30 healthy volunteers studied the effect of oesophageal function (assessed by barium swallow) on intake (4 days), anthropometric recordings and biochemical parameters (trace element and fatty acids) . Dysmotility was demonstrated in 93%, and although some were asymptomatic the majority had one or more symptoms: dysphagia (67%), regurgitation (50%), vomiting (20%). No difference in biochemical parameters was found compared with healthy volunteers. Dietary questionnaires identified low fibre intake in SSc, but no energy or macronutrient differences, although subanalysis suggested increased fat intake with dysphagia. However, this study may be limited by the small number of symptomatic patients.
Additionally, the CSRG study showed no association between dysphagia or reflux and malnutrition risk, as assessed by the Malnutrition Universal Screening Tool (MUST) score (Fig. 1) . MUST was developed by the British Association of Enteral and Parenteral Nutrition (BAPEN) to screen for malnutrition risk . It scores malnutrition risk using assessments of BMI, recent unplanned weight loss and acute disease effect . It is therefore a relatively poor tool for chronic and gradually progressive disorders. Of 586 patients, 54% reported swallowing difficulty, 64% reflux and 43% retrosternal burning. Adjusted odds ratios for associations between swallowing difficulty, reflux and retrosternal burning with MUST were non-significant . This may reflect the ungraded nature of symptom severity and the limitations of MUST in chronic disease.
Gastric dysmotility occurs in both lcSSc and dSSc, but there are relatively few studies evaluating gastric involvement. In a gastric emptying study conducted using a radiolabelled solid meal, 50% of 20 symptomatic patients fulfilled absolute criteria for delayed emptying (>2 s.d.), with 63% having some delay . Gastric dysfunction appears to result from autonomic dysfunction and fibrotic change of the muscularis mucosa and submucosa [37, 38]. Altered gut hormone profiles are also implicated. In SSc, plasma vasoactive intestinal peptide and motilin levels correlate with coordinated gastric contractions . However, causality has not been studied.
Gastroparesis leads to early satiety, nausea and vomiting . Prokinetics are often prescribed despite only a few small studies showing benefit in SSc (metoclopramide n = 4, cisapride n = 9) [36, 41]. The CSRG reported that early satiety was significantly associated with malnutrition risk on bivariate analysis . However, whether gastric dysmotility leads to nutritional decline in SSc has yet to be prospectively examined.
Small intestinal manifestations
Small intestinal dysmotility is common in SSc, with 88% of 17 unselected patients (35% with GI symptoms) displaying aberrant small bowel manometry in a prospective study . Radiological contrast studies demonstrate typical features of a megaduodenum, small bowel dilation, often with diverticulae, and the typical hide-bound appearance (42%; Fig. 2). Pneumatosis intestinalis is relatively uncommon (8%) but its presence should indicate a possible diagnosis of SSc . Dysmotility and stasis frequently predispose to small intestinal bacterial overgrowth; glucose hydrogen and methane breath testing demonstrated bacterial overgrowth in 43% of 51 unselected patients (no difference between dSSc and lcSSc) [44, 45]. Immunosuppressive use did not increase risk . Prevalences determined by breath testing alone may be underestimates, as the microbiological testing of jejunal aspirates identified additional cases compared with breath testing .
Bacterial overgrowth leads to nutrient malabsorption. The CSRG detected, on bivariate analysis, significant associations between MUST and diarrhoea, necessitating antibiotics . In difficult cases, prolonged cyclical antibiotic therapy may be needed. Somatostatin analogues may improve motility, but evidence is limited, and use may be associated with perforation [47–49]. Artificial feeding, including PN, may become necessary in severe refractory cases .
SSc may involve the colon and anorectum; radionuclide colonic transit studies in 23 unselected patients (dSSc and lcSSc) showed delayed transit in 57%, compared with 20 age-matched healthy control subjects . Colonic involvement produces symptomatic constipation and, in its severest form, results in pseudo-obstruction. A symptom review in 401 consecutive patients (SSc-GIT 1.0 questionnaire) identified that approximately 45% experienced weekly constipation . The CSRG reported an association between MUST and patient-reported constipation . Although colonic motility has limited relevance to nutrition, it is usually associated with small bowel dysmotility. Marked postprandial symptoms may be severe enough to impair eating and, in the severest pseudo-obstruction cases, PN may be necessary .
Primary biliary cirrhosis (PBC) is a chronic progressive cholestatic hepatic disorder with increased prevalence in SSc, especially lcSSc. PBC-specific autoantibodies have been detected in 15% (62% lcSSc) of 52 consecutive patients (63% lcSSc) . However, in a series of 817 patients (60% lcSSc), only 2% (81% lcSSc) had confirmed PBC, with compatible histology or abnormal liver enzymes . This is still greater than in the general population (<0.02%) . Resultant cholestasis is associated with fat malabsorption and fat-soluble vitamin deficiency . Bacterial overgrowth probably contributes more to this deficiency syndrome. The incidence of malnutrition in patients with concomitant PBC and SSc has not been evaluated.
Exocrine pancreatic insufficiency is a feature of dSSc. In a series of 16 consecutive patients (dSSc), 7 had biochemically proven reduced exocrine function, of which 4 were clinically relevant . The pathophysiology is poorly understood, although histology from a case of pancreatic necrosis reported arterial vascular changes and focal fibrosis of the interstium . Untreated pancreatic insufficiency, unrelated to SSc, causes malabsorption and results in weight loss. However, there are no reports of associated nutritional compromise in SSc. The CSRG attempted to address this . They sought a link between malnutrition risk and steatorrhoea, a possible symptom of untreated pancreatic insufficiency and/or bacterial overgrowth. No association was shown but lack of symptom specificity may have influenced the result .
Cardiovascular and pulmonary complications are common. Cardiac manifestations include conduction defects, ischemia and, ultimately, heart failure, which may also be due to pulmonary artery hypertension [58–60]. Cardiac involvement is responsible for 26% of deaths and pulmonary hypertension for 26% . Pulmonary involvement produces interstitial fibrosis in 19–47% and death in 35% [59, 61].
Studies in non-SSc patient populations with similar cardiovascular and respiratory pathology demonstrate malnutrition. Dietary studies in patients with chronic heart failure (n = 39) showed that 73% were not meeting recommended energy intakes . Cachexia may complicate chronic heart failure and worsen prognosis . Up to 27% of patients with idiopathic interstitial fibrosis referred to transplantation are malnourished . No studies have assessed the nutritional impact from cardiovascular or respiratory involvement in SSc.
Inflammation, anorexia and gut hormones
An association between pro-inflammatory cytokines and malnutrition in SSc has not yet been evaluated. However, inflammation is associated with malnutrition in other extra-gastrointestinal conditions, including end-stage renal disease, chronic heart failure, cancers and RA [65–68]. However, the molecular signals linking inflammation to anorexia and impaired energy homeostasis are unclear.
Normal GI sensations and abnormal symptoms are determined by the interplay between gut peptides and sensory afferent neurones inputting to the enteric and autonomic nervous system [69, 70]. Since patients with SSc often display features of autonomic nervous dysfunction, it is postulated that aberrant gut peptide responses and autonomic dysfunction are central to the development of GI symptoms such as early satiation and anorexia [71, 72]. This is supported by anorexia in other inflammatory GI disorders being associated with the up-regulation of satiety hormones, such as cholecystokinin, glucagon-like peptide-1 and peptide-YY . Furthermore, aberrant peptide responses have been implicated in the pathogenesis of GI dysfunction in SSc, although further research is required to confirm and delineate the mechanistic pathways [39, 74]. GI peptides are an unexplored area of SSc research that may offer insight into the pathogenesis of patients' anorexia and GI dysfunction, and possible therapeutic targets.
When the CSRG studied 586 patients with SSc, almost 18% had high malnutrition risk (MUST) . As described earlier in this article, many processes contribute, in varying degrees. The importance of screening for, and managing, malnutrition in SSc has been identified by a North American panel of experts .
In the UK, the National Institute for Health and Clinical Excellence (NICE) advocates nutritional screening of all in-patients and first out-patient attendees (except identified low-risk groups) . The American Society of Parenteral and Enteral Nutrition (ASPEN) and the European Society of Parenteral and Enteral Nutrition (ESPEN) recommend screening all patients on admission [77, 78]. ESPEN, NICE and the North American expert panel recommend MUST [6, 76, 78].
MUST is a validated screening tool for malnutrition risk, irrespective of underlying disease (Fig. 1) . It is the only screening tool to have been used in SSc [6, 80]. MUST scores BMI, percentage unplanned weight loss in the preceding 3–6 months and acute disease effect. Each may score up to 2 points. The sum of these represents risk of malnutrition: 0 (no risk), 1 (medium risk), 2 or more (high risk) [81, 82]. Further information about MUST is available from the BAPEN website (www.bapen.org.uk). Other studied assessment tools, including BMI, food diaries, anthropometry and bioelectrical impedance have been compared with MUST, with variable results.
It is crucial to note that serum albumin is not a nutritional marker, since levels are confounded by many variables, such as sepsis and inflammation, particularly in the acute phase . It is therefore unsurprising that when serum albumin was evaluated in patients with a 10-year average SSc disease duration, only 2% had subnormal albumin despite 21% having high malnutrition risk (MUST) .
A study investigating early mortality predictors in 250 patients with SSc identified a strong association with BMI <18.5 . However, BMI is limited as a standalone screening tool and needs to be assessed in the dynamic context of MUST. Indeed, BMI is not validated as a lone nutritional screening tool in SSc [6, 7]. The CSRG found similar BMI distributions in patients with SSc and the general population, despite almost 18% of patients having high malnutrition risk . Another study involving 124 patients compared BMI to bioelectrical impedance, with no significant correlation .
Food diaries are routinely used in dietetic assessments. Only two studies have evaluated their use in SSc. One compared 3-day dietary histories between 30 patients and healthy volunteers, and found no difference in energy or macronutrient intake . Only fibre was lower in SSc. The other compared specific micronutrient intake between 12 patients and healthy volunteers . Intake was not the cause of the reduced serum antioxidant levels in the SSc group. However, interpretation of these studies is limited by sample size, end-point and study duration.
Anthropometrics allow objective assessment of nutritional status and have functional relevance. Individual measurements (BMI, arm circumference and triceps skin-fold thickness) are plotted against standardized population tables for health status assessment. In 30 patients with SSc (17 dSSc, 13 lcSSc), triceps skin-fold thickness, mid-arm circumference and arm muscle circumference were measured and compared with healthy volunteers . Mean triceps skin-fold thickness was significantly higher in female patients than in volunteers and non-significantly higher in males. Arm circumferences were significantly lower in male and female patients than in volunteers, despite BMI being higher between the female groups and not significantly lower between the male groups. These differing results may be due to the confounding effect of increased skin thickness in SSc. Only two patients had both a triceps skin-fold thickness and mid-arm circumference below the 5th centile. Both required nutritional support, suggesting that concordance of very low measurements may predict malnutrition.
Bioelectrical impedance differentiates lean and fat mass through their differing conductivities. Results convey nutritional status and hydration. Results from a bioelectrical impedance study in SSc (n = 124) were reported to reflect malnutrition .
To the best of our knowledge, no observational studies document the rate of nutritional decline in SSc. MUST, which is weighted towards acute illness and rapid weight loss, may have limited use in chronic, slowly progressive disorders. Therefore, no evidence exists on which to base recommendations for timing of nutritional screening. However, our minimum recommendation, in the absence of such evidence, is for screening at diagnosis and annual review. An expert panel recommends that review should include the basic investigations of haemoglobin (haematinic deficiency), serum folate (bacterial overgrowth) and carotene (fat malabsorption) . However, carotene and folate’s reliability as screening tools for either manifestation remains unproved, and in some studies haemoglobin has proved unreliable .
Patients identified at risk of malnutrition by any modality need appropriate dietetic assessment and intervention . Expert physicians are essential to assess and manage contributing GI and systemic disease manifestations. Specific pharmacological intervention could prevent avoidable complications (e.g. oesophageal strictures, bacterial overgrowth) and improve contributory symptoms (early satiety, dysphagia, constipation, digital ulcer pain) [17, 32, 45, 85, 86]. Occupational therapist input may benefit patients by reducing functional barriers. However, a Canadian postal survey of over 900 patients with SSc found that only a third had been assessed by an occupational therapist . Benefit from occupational therapy has been demonstrated in other chronic illnesses but has yet to be investigated in SSc . Dentists are needed for good oral health, as education and regular oral care reduce caries [24, 89]. As tooth loss is associated with poor nutrition, and dentures are problematic with microstomia, proactive preventive dentistry may have nutritional benefits, although this has yet to be confirmed .
Nutritional intervention: approaches
Nutritional interventions take many forms. They include both general approaches, such as those described earlier, and specific interventions. Specific interventions target intake through the provision of dietary counselling and artificial support (enteral and parenteral).
Nutritional intervention: dietary counselling
Although dietetic assessment is advocated in response to high MUST scores, no studies have investigated dietetic counselling in SSc. The benefits are, however, supported by evidence from other chronic conditions [90, 91]. As yet, no specific recommendations regarding dietary interventions exist for patients with SSc. General recommendations can, however, be tailored according to symptoms and disease pattern; for example, frequent, small, low-fibre and low-fat meals with increased liquid nutrient intake are recommended for gastroparesis .
Nutritional intervention: artificial nutritional support
Current recommendations for providing adult nutritional support advise the use of oral, enteral or parenteral support, either alone or in combination . Enteral tube feeding should be instigated when requirements cannot be met by oral intake alone, and PN instigated when oral or enteral feeding are not sufficient, possible or appropriate .
Enteral tube feeding involves the delivery of liquid feed through a tube into the GI tract. In the first instance, the potential benefit and tolerance of this should be explored by conducting a nasoenteric trial. A trial involves either nasogastric (NG) or nasojejunal (NJ) fine-bore tube feeding for a short pre-determined period. To date, no studies have determined an optimum period. The NJ route has potential benefits over the NG route, by delivering post-pyloric feed. This is beneficial in the presence of gastroparesis. If feeding is tolerated and nutritional benefit is evident, and it is determined that the patient will require long-term artificial nutritional support, then placement of a secure percutaneous enteral feeding tube is appropriate.
Several different types of permanent percutaneous tubes exist. These include a percutaneous endoscopic gastrostomy (PEG), percutaneous endoscopic jejunostomy (PEJ), percutaneous endoscopic gastrostomy with jejunal extension (PEG-J) or direct surgical jejunostomy. They differ in their insertion techniques and feed delivery points. Only the PEG delivers food into the stomach, while the others feed into the jejunum. No studies have investigated artificial tube feeding, via any route, in SSc, although small case series exist . In our experience, if there is any suggestion of significant gastric dysmotility, an NJ trial followed by jejunostomy placement, in conjunction with prokinetics, postural advice and any necessary therapy for bacterial overgrowth, is appropriate. As yet, no studies report jejunostomy feeding in SSc, despite its proven benefit in other causes of severe refractory gastroparesis . Jejunal feeding has the added benefit of reducing the associated aspiration risk, which is clearly important in patients with concomitant lung disease .
PN is necessary when oral and/or enteral nutrition are inadequate or cannot be tolerated. The decision to proceed to PN should be made in conjunction with, and subsequently managed by, an experienced nutrition team because of the potentially serious complications (e.g. catheter-related sepsis, thrombosis), which may arise due to the complex nature of PN. With specialist input, PN may either provide temporary support, while other feeding modalities are established, or long-term nutrition, when all other options have failed. Long-term PN is managed at home by patients, supported by specialist nutrition nurses, and is termed home parenteral nutrition (HPN).
The first report of HPN use for SSc was in 1981 . Since then, only four small series detailing HPN experience have been published, although patients have been included in non-specific HPN series. The four specific series involved only 15, 4, 8 and 12 patients respectively [8, 51, 96, 97]. All reported severe GI tract involvement as an indication. The series published in 1999 described four patients who required HPN for intestinal pseudo-obstruction; the high sepsis rate of 50% is notable . The earliest series (1989) reported a single centre’s experience over 8 years (1979–1987), incorporating 15 700 catheter-days . During this time, 15 patients commenced PN and continued for a mean of 2.9 years (range 2 months to 7.5 years). Three patients' treatment was subsequently discontinued. Subjective assessment of improved quality of life was reported by 11 patients. Unfortunately, seven patients died, but none of these deaths was directly related to PN. There were two cases of septicaemia and two catheter site infections.
The SSc-specific series from our unit reported the outcomes of eight patients who received PN between 1993 and 2006 . The median PN period was 40 months (0.8–192 months) and covered 13 851 catheter-days. The main indication was ongoing nutritional decline, despite oral and enteral support (PEG in three patients). In this series, three patients died of causes unrelated to HPN, while four were still on PN at the end of the study. Of these four patients, two had >100 months of HPN (103 and 192), suggesting that HPN prolongs life. Complication rates were comparable to those from PN given for non-SSc intestinal failure. There were two line infections, giving a line infection rate of 0.14 per 1000 catheter-days, which compared with a rate of 0.52 per 1000 catheter-days in patients on HPN without SSc. Other complications included line occlusions (eight blocked, four thrombus), line calcification (one) and fluid overload. Training was given to seven patients (one died within 1 month) but, ultimately, five were unable to complete the training due to functional limitations and required support from family (four patients) or district nurse (one patient). Dependence on others for line care has long-term financial, emotional and social implications for both patient and caregiver .
The most recent series of 15 patients (1998–2010) reported a 58% 2-year survival on PN, though causes of death were not PN related but were instead due to complications of SSc . This compares with a reported 82.9% 5-year survival of non-oncological HPN patients . This highlights a worse prognosis for patients with SSc requiring HPN, presumably due to the multi-systemic nature of SSc.
In summary, from the small published series HPN appears to represent a viable option for maintaining nutrition in SSc when other approaches fail. Patients with SSc appear to have greater dependence on others for line care. However, complications are similar to those in patients without SSc who require HPN. Larger patient series are required to clarify this point.
Summary and future perspectives
Despite the high prevalence of malnutrition in SSc and its associated morbidity, much about it remains poorly understood. There are a number of possible contributing factors, which may have variable degrees of influence in different patients. These contributing factors include gastrointestinal manifestations, which are present in approximately 90% of patients, mood disorders, functional influences and other systemic disease effects (including cardiac and respiratory involvement). All of these require further study to fully understand their pathophysiology, natural history and progression. Only by fully understanding the nature of each disease manifestation in SSc can the future management of the disease be optimized and nutritional complications prevented.
Little is known about the nature of nutritional decline in SSc, including its rate of progression and impact. Such knowledge could enable early detection—and, possibly, avoidance—of malnutrition. However, few studies have actually investigated nutritional interventions, such as dietetic advice and artificial nutritional support, in SSc. Therefore, many of the interventions used in the nutritional management of SSc are derived from experience of managing nutritional impairment in other diseases.
We acknowledge Dr Neal Townsend (Consultant Radiologist, Salford Royal NHS Foundation Trust) for his assistance in the acquisition of the CT image.
Funding: This work was supported by the Raynaud’s and Scleroderma Association.
Disclosure statement: E.H. has received a research grant from the Raynaud’s and Scleroderma Association. S.L. has received educational support and served as a speaker and an advisory board member for IBD/nutrition companies including Baxter, Ferring, Shire and MSD and has received research funding from Crohn’s Colitis UK and the Raynaud’s and Scleroderma Association. All other authors have declared no conflicts of interest.