Epidemiology and hospitalizations

Gastroparesis is a disabling gastrointestinal motility disorder that can significantly impact patient lives. The true prevalence of gastroparesis is unknown and difficult to determine for several reasons: (1) there is a poor correlation between gastrointestinal symptoms and the rate of gastric emptying; (2) it is unclear how many patients with nonspecific symptoms of gastroparesis actually seek healthcare and/or are referred to a gastroenterologist, and (3) scintigraphy remains the gold standard for the measurement of gastric emptying, a relatively expensive test associated with a small amount of radiation exposure slowly expanding from specialized centers to community hospitals . It is estimated that at least 5 million people, and conceivably up to 10 million people in the US are affected by gastroparesis . In one of the largest epidemiologic population-based studies on gastroparesis, the age-adjusted prevalence of definite gastroparesis per 100,000 persons was 37.8 in women (95% CI, 23.3–52.4) and 9.6 in men (95% CI, 1.8–17.4) .

A notable trend observed over the past couple of decades includes a significant increase in gastroparesis-related hospitalizations across the United States (U.S.). Reasons contributing to hospitalizations include poor glycemic control, infections, and noncompliance with or intolerance of medications . In a retrospective study of 326 patients, more than one-third of gastroparetic patients required hospitalizations and emergency room visits with diabetics requiring the highest number of hospitalization days . A hallmark study by Wang et al. that examined hospitalization trends related to gastroparesis in the U.S. between 1995 and 2004 showed a 138% increase in hospitalizations with a major increase occurring after the year 2000 . Data was collected from the Nationwide Inpatient Sample, the largest inpatient care database publicly available in the U.S comprising a nationally representative sample of 5–8 million hospitalizations per year from about 1000 hospitals each year. When further broken down, hospitalizations with gastroparesis as the primary diagnosis increased by 158% (from 3977 to 10,252), and hospitalizations with gastroparesis as the secondary diagnosis increased by 136% (from 56,726 to 134,146). Five upper gastrointestinal (GI) conditions were studied including gastroesophageal reflux disease (GERD), gastric ulcers, gastritis, and nonspecific nausea/vomiting. Diabetes was listed as a comorbidity for 21% of patients with gastroparesis in 1995, versus 26.7% in 2004 suggesting an increasing prevalence of diabetes. Of all five upper GI conditions studied, gastroparesis had the longest hospital length of stay (+15.4% to +66.2%, all P <.001). Gastroparesis also had the highest medical charges in 2004 ($20,573 for primary gastroparesis versus $24,965 for secondary gastroparesis).

Reasons for increasing gastroparesis-related hospitalizations in the U.S. are unclear but the authors suggest several theories: improved recognition and/or diagnosis of gastroparesis, increased prevalence of diabetes, increasing prevalence and severity of gastroparesis and/or changes in the diagnostic criteria and treatment of gastroparesis. One limitation to this hospitalization data was that the etiology of gastroparesis (i.e. idiopathic, diabetes, post-surgical) could not be ascertained as past surgical history was not available . The authors further present several personal hypotheses as to why there may have been a striking increase in gastroparesis-related hospitalizations after the year 2000. First, the withdrawal of the gastroprokinetic agent cisapride (Propulsid) from the U.S. market in July of 2000 may have led to more symptomatic patients now requiring hospital care. Second, the approval of the gastric electrical neurostimulator (Enterra® device) by the FDA through a Humanitarian Device Exemption occurred in 2000, and surgical implantation of this device requires hospitalization. Thirdly, the 4-hour gastric emptying scintigraphic study was published in 2000 and this methodology increased the diagnostic yield of gastroparesis .

A fourth personal theory as to why there may have been significantly more hospitalizations over the last couple of decades is that cyclic vomiting syndrome (CVS) was not as recognized as it is today, and patients with chronic nausea and vomiting, who in retrospect could have been CVS patients, were being admitted and incorrectly labeled with a diagnosis of suspected idiopathic gastroparesis attributed to a post-infectious gastroenteritis . Patients invariably received narcotics in the emergency department or soon after hospitalization which slowed their gastric emptying so when the scintigraphic gastric emptying study was performed within a day or so of admission, and was delayed, then this was interpreted ‘gastroparesis’ and the patient was thus labeled as a “gastroparetic” for life. A recent study by McCallum et al. showed that up to 10% of patients referred to a major medical center for evaluation of nausea and vomiting and diagnosed as gastroparesis due to the circumstances described above actually had CVS . This figure is definitely underestimated given the evolving picture of “cannabis hyperemesis” that we are now also beginning to fully appreciate.

Another report by Nusrat and Bielfeldt found that annual hospitalizations for gastroparesis increased more than 18-fold from 918 to 17,736 between 1994 and 2009 . The authors attributed this surge in hospitalizations to an increased awareness of the disease and/or classification rather than a true difference in the incidence and/or prevalence of gastroparesis since there was also a concomitant decrease in resource utilization for other functional gastrointestinal disorders. Comparable findings were found in the important large population-based cohort study from Olmstead County, Minnesota in that about 25% of their patients with gastroparesis required therapeutic interventions such as enteral feeding which required hospitalization . Wang et al. suggested there may be an increasing prevalence of gastroparesis in the U.S. which puts a burden on our health care system and also has growing economic impact.

These studies collectively show that frequent hospitalizations are associated with substantial morbidity and greatly impair the quality of life in patients with gastroparesis, notwithstanding the substantial economic burden and impact it places on health-care resources. Furthermore, hospitalizations are correlated with extensive and repeat diagnostic testing which further contribute to healthcare costs and risk of care with limited impact on overall treatment and outcome.

Clinical course

Several long-term longitudinal studies have aimed to evaluate the natural history and prognosis in patients with gastroparesis. In a cohort study of 86 patients with diabetic gastroparesis followed over 9 years, approximately 25% died by the end of the study although they found no evidence that a delay in gastric emptying was associated with higher risk of death after adjustment of comorbidities . Patients with long-standing diabetes are known to have many end-stage complications that further contribute to their morbidity and mortality – beyond a diagnosis of gastroparesis. Another longitudinal study aimed to evaluate the natural history in 20 diabetic patients followed over 12 years and surprisingly showed no significant changes in gastric emptying and upper gastrointestinal symptoms over time . However, this particular study had a very specific population sample since it was performed in an academic endocrinology unit managing diabetics and may not reflect a GI referral-type of population. Other studies show that gastroparesis is associated with significant morbidity and poor prognosis with increased hospitalizations, emergency room visits, and decreased quality of life. In a tertiary referral center study where 146 patients were followed over six years, 7% died, 21% required long-term parenteral or enteral feeding, 5% had a gastrectomy, and 6.2% had placement of a gastric electrical stimulator . In the landmark population based study from Olmstead County in Minnesota over a 10-year period, one-third of patients died and another one-third required hospitalization, tube feeding or medications related to gastroparesis . It was observed that the overall survival in patients with gastroparesis was significantly lower than that of the Minnesota population after adjustment for gender and age (67% vs 81%, P <.01).

A large cohort study of 361 patients enrolled in the GpCRC and followed prospectively over a 48 week period showed that gastroparesis is not a benign condition and the overall burden of this disease remains high . Ultimately, 6.5% of patients required TPN, 5% required J-tube feeding, and 30.5% required gastric electrical stimulation. Hospitalizations did not significantly change during follow-up. Approximately 45% required hospital admission. At the end of the 48 week period, quality of life showed mild improvement in total scores by the Patient Assessment of Upper Gastrointestinal Symptoms (PAGI-QOL) (2.9 vs 2.5; P =.02), and the Gastroparesis Cardinal Symptom Index (GCSI). (3.1 vs 2.8; P <.001) . The PAGI-QOL questionnaire is a validated tool that measures quality of life in patients with upper gastrointestinal disorders (dyspepsia, gastroparesis, gastroesophageal reflux) with scores ranging from 0 to 5 (higher scores reflect improved QOL). The GCSI questionnaire is a reliable and well-accepted tool to assess symptom-severity in gastroparetic patients with scores ranging from 0 to 5 (higher scores reflect worse symptoms). Despite mild improvement in these clinical parameters, the authors concluded that the overall morbidity and health impact of gastroparesis remains high, even during treatment at leading academic centers using all pharmacologic agents available. Hence an obvious conclusion is that novel therapeutic approaches or perhaps endoscopic or surgical interventions in the future are needed to improve the symptoms and quality of life in this patient population. Another example from 146 gastroparetic patients in an academic tertiary referral center followed over six years showed that at the end of the follow-up period, 74% required continuous prokinetic therapy, 22% discontinued prokinetics, 5% underwent gastrectomy, 6.2% had implantation of a gastric electrical stimulator, and 7% passed away . During this six year period, at some point 21% required enteral nutrition with a feeding jejunostomy tube or periods of parenteral nutrition.

Another study analyzed morbidity, mortality, and predictors of improvement in 358 gastroparetic patients enrolled in the GpCRC and followed over 4 years receiving the best care available at leading GI motility academic centers . 70% of patients had IG, 30% had diabetic gastroparesis (DG), 73% of patients had delayed gastric emptying (GE), and 27% with chronic unexplained nausea and vomiting had a normal GE. Over the 4-year course of follow-up, GE status (delayed or normal) ( P =.75), was not associated with improvement in symptoms ( P =.75). About one-third of all patients with delayed GE experienced improvement in symptoms (symptom improvement was defined by a >1 point decline in GCSI score). 5% of patients died. This incident death rate (IDR) was higher in patients with DG compared to IG and in those with delayed GE compared to normal GE. Factors associated with clinical improvement included normal BMI, less depression/anxiety, initial infectious prodrome, older age, and less abdominal pain. The authors observed that most patients with GP do not improve over time .

There are hints of more positive outcomes with several long-term treatment studies with gastric electrical stimulators (GES). The longest and largest single center study ever performed to date to assess the clinical outcomes on GES therapy was performed by McCallum et al. . They reported on a large cohort of 188 gastroparetic patients with a mean follow up of 56 months. Patients were treated with Enterra GES therapy which was approved by the US Food and Drug Administration (FDA) as a Humanitarian Device Exemption in 2000 for treating symptoms of DG and IG refractory to all medical management. Results did show sustained efficacy and clinical improvement up to 10 years for 54% of patients who experienced a reduction of upper GI symptoms of approximately 50% . Diabetic and postsurgical gastroparetic patients achieved a greater degree of clinical improvement compared to those with IG. These results were consistent with other studies reported from their center , as well as with other studies that also assessed the efficacy of GES in patients with intractable gastroparesis . Other notable findings in the McCallum et al. study assessing clinical outcomes of GES therapy included decreased hospitalizations, reduction of HbA1c level, improved quality of life, decreased prokinetic and/or antiemetics use, and weight gain with eventual removal of 89% of jejunal-tubes originally placed for enteral nutrition supplementation.

What do we know about mortality in patients with gastroparesis?

Causes of death in patients with gastroparesis can be attributed to the following: (1) complications from underlying causes of gastroparesis (2), treatment-related complications (3), gastroparesis-related complications (i.e. nutritional and metabolic consequences) . Unfortunately, the mortality of gastroparesis is not well defined and conflicting reports exist in the literature. In a study of 146 gastroparetic patients followed over six years in an academic referral center 7% died by the end of the follow-up period . Other data available from case series and tertiary referral centers show that mortality rates in gastroparesis patients can range between 4% and 38% . However, some argue that data obtained from these tertiary referral centers reflect a skewed patient population of the worse cases of gastroparesis and that all gastroparetic patients do not necessarily have this high mortality risk . In a prospective cohort study of 86 diabetic patients, Kong et al. showed that gastroparesis is not associated with a major increased risk of mortality with similar mortality noted in patients with DG compared to diabetic patients without gastroparesis . In this study, the median time of death in a patient with DG was 6 years (range 1–12) from time of diagnosis, and mortality was mostly associated with complications from micro and macroangiopathic end-organ damage and not directly due to gastroparesis . The majority of these patients were alive after a period of 9–14 years and the risk of death was not related to the rate of either liquid or solid gastric emptying . Later, the same authors confirmed these findings in one of the longest ever published longitudinal studies aimed to evaluate the mortality and prognosis of diabetic gastroparesis in these same 86 diabetic patients followed over 25 years . By the end of the study, about one-third died with cardiovascular disease being the most common cause of death. Even though delayed gastric emptying was observed in slightly less than half of these patients, gastroparesis was not deemed to be a factor associated with mortality though advanced age and autonomic nerve dysfunction were.

Other studies have concluded to the contrary suggesting that gastroparesis is, in fact, associated with greater mortality, especially diabetic patients with more symptomatic gastroparesis . Cutts et al. showed a higher mortality in severely symptomatic gastroparetic patients who were referred for a gastric electrical stimulator (GES) but did not undergo placement of GES . However, in the Olmstead County epidemiologic study, one-third of all incident cases of gastroparesis patients died, with the most common causes of death attributed to cardiovascular disease . And while this large population-based study confirmed mortality in gastroparesis, this was mostly due to an increased mortality in diabetics who had comorbidities. This theory is further supported by the data analysis performed by Bielefeldt using the Nationwide Inpatient Sample (NIS) and Nationwide Emergency Department Sample (NEDS) databases over a four year period which showed that gastroparesis is not related to a high mortality risk, with death primarily occurring because of complications from comorbidities, rather than as a consequence of gastroparesis .

These latter studies are important because they change the perspective on how we view the natural history of gastroparesis. It is reassuring to know that even though gastroparesis significantly affects quality of life, it may not “come with the burden of limiting ‘quantity’ of life” as stated by Bielefeldt after he reviewed the 25-year study by Chang et al. that failed to show that DG is associated with increased mortality or poor prognosis. In short, we need more long-term population-based studies that better define the prognosis and clinical mortality of gastroparesis.

Long term follow-up of gastroparesis: does gastric emptying change over time?

The gold standard for measuring gastric emptying is scintigraphy based on consensus recommendations from the American Neurogastroenterology and Motility Society and the Society of Nuclear Medicine . A low-fat, egg-white meal as described by Tougas et al. followed by imaging at 0, 1, 2, and 4 hours after meal ingestion provides standardized information about normal and delayed GE . Gastric retention of >90% at 1 hour, >60% at 2 hours, and >10% at 4 hours is indicative of delayed emptying. While delayed gastric emptying is the clinical feature in gastroparesis, some studies have shown a weak association between the severity of gastrointestinal symptoms and the rate of gastric emptying .

Several studies discuss the relationship between changes in gastric emptying over time in patients with gastroparesis. Jung et al. demonstrated that having delayed gastric emptying in diabetics with symptoms of gastroparesis predicted morbidity with increased hospitalizations and emergency room visits . Preliminary data from the GpCRC assessed gastric emptying (GE) changes over 48 weeks and showed that GE is more likely to normalize or improve in patients with IG compared to DG (41% vs 22%, P =.07) . Patients with DG had worsening GE changes over 48 weeks compared to patients with IG (25% vs 12%, P =.06). Improved GE was associated with improved symptoms (nausea, upper abdominal discomfort) and increased weight. Worsened GE was associated with a longer duration of symptoms ( P <.01), lower SF-36 physical quality of life score ( P =.08), and having a gastric electric stimulator ( P =.05). A take-home point from this study which should be considered when counseling different subsets of gastroparesis patients is that IG patients are more likely to improve and even normalize over a 48 week period, while GE in DG patients will not improve and invariably may get worse with time.

A recent study by Hasler et al. assessed long-term symptom and quality of life outcomes in patients with suspected gastroparesis from diabetic versus nondiabetic etiologies after gastric emptying testing with a gastric emptying study (GES) and wireless motility capsule (WMC) . This study enrolled 150 patients from 10 different centers with gastroparesis symptoms and measured symptoms and quality of life (QOL) at baseline and then at 6 months, which were then compared in non-diabetics (n=114) and diabetics (n=53—18 T1DM, 35 T2DM). Symptoms and QOL were measured by the GCSI, PAGI-SYM, (0=none, 5=very severe for each symptom item) and PAGI-QOL (0=poor, 5=excellent). In a nutshell, both symptoms and QOL worsened over 6 months for diabetic patients with abnormal emptying, whereas both symptoms and QOL improved in non-diabetics. Only QOL improved in patients with T2DM with normal emptying but not in T1DM with gastroparesis. The take-home point from this study was that having abnormal GE in diabetics was associated with a worse QOL and no improvement in symptoms—whereas having normal GE in diabetics was associated with improved QOL and improvement in all symptoms (except nausea, vomiting) with treatment (drug and diet therapy) and time 6 months into this study. One could speculate that glucose control will vary greatly in diabetics over a 6-month time frame and hyperglycemic events induce nausea and vomiting. Moreover, hyperglycemia in diabetics is further exacerbated by infections, non-compliance to medications, and “stress” – and over 6 months, infections likely occurred, particularly urinary tract infections .

The disparities between GE and symptoms identified in the above study by Hasler et al. are also evidenced in other reports. Loss of interstitial cells of Cajal (ICC) correlate with delayed gastric emptying in DG with greater loss of ICC associated with slower GE . However, the loss of both ICC and enteric nerves did not correlate with symptom severity – just as gastric emptying results do not always correlate well with symptom severity. Also, patients with diabetes with a documented GE delay and symptoms of gastroparesis were found to have a higher likelihood of concomitant microangiopathies and macroangiopathies such as cardiovascular disease, hypertension, and retinopathy . This suggests that delayed gastric emptying in diabetic patients predicts negative health outcomes. Additionally, the duration of diabetes and the presence of secondary complications seem to be most predictive of poor clinical outcomes – not a documented delay in gastric emptying.

To further emphasize the point that changes in GE do not necessarily correlate with symptom severity in patients with gastroparesis, we will now discuss long-term changes in GE after gastric electrical stimulation (GES) therapy. As previously mentioned – in the 10 year study by McCallum et al. assessing long-term clinical outcomes of GES therapy in 221 patients with severe gastroparesis (142 diabetic, 48 idiopathic, 31 postsurgical) – 26% of the 119 patients with GE follow-up data normalized their results . However, the overall results were poor with the 2-hour gastric retention decreasing from a median of 70% at baseline to 66% (>60% retention at 2 hours is abnormal) and 4-hour retention decreasing from 37% to 30% at time of last follow-up (>10% retention at 4 hours is abnormal). Nevertheless, there was sustained improvement in these patients of approximately 50% reduction in symptoms for up to 10 years, particularly reduction in nausea and vomiting.

Our group at Texas Tech University El Paso recently investigated the long term follow up of IG patients and the relationship between changes in GE and GI symptoms . After 1 year of treatment with antiemetics and prokinetics, nearly 60% of 23 IG patients normalized their GE compared to their baseline GES at time of diagnosis. Patients who normalized their GE also had improvement in their Patient Assessment of Upper Gastrointestinal Symptoms (PAGI-SYM) mean score from 44.61 at baseline to 35.6 one year later. However, patients who normalized their gastric emptying still had persistent symptoms.

Subgroup differences