Demographics of Gastroparesis Hospitalizations Through the Age Spectrum Using National Inpatient Databases

Children Compared With Adults

Abstract

Goal: 

The goal of this study was to characterize the etiology and demographics of hospitalized patients with gastroparesis (Gp) across different age groups.

Background: 

Gp is a chronic condition associated with increasing hospitalizations and costs. The gender and etiology distributions of Gp throughout the age spectrum are unknown.

Materials and Methods: 

Nationwide Inpatient Sample (NIS) and Kid’s Inpatient Database (KID) were used to identify patients using International Classification of Diseases (ICD)-10 codes for Gp as a primary diagnosis or as secondary diagnosis with the first diagnosis a GI-related symptom.

Results: 

There were a total of 15,790 admissions (75.6% female, age: 46.2±18.0 y). After age 6, female admissions percentage increased: ages 2 to 5: 45.0%, ages 6 to 12: 62.8%, ages 13 to 20: 76.7% (P<0.001), with a distinct increase at age 12. Diabetic gastroparesis (DG) was seen in 3995 (25.3%) of all Gp admissions but in only 1.1% of children under the age of 20. Overall, 68% of DG admissions were female, but a higher percentage of DG was seen among male admissions for Gp compared with female admissions for Gp between ages 21 and 64 (38.3% vs. 23%, P<0.001). The most common races were white (63.2%), African American (20.6%), and Hispanic (8.7%). DG was more often present in Native American (61.9%), Hispanic (39.1%), and African American (38.2%) admissions than in white patients (17.8%; P<0.05).

Conclusions: 

This study using large inpatient databases shows that the gender, race, and etiology of Gp admissions is age-dependent. The female predominance of Gp admissions is more prominent from the second decade of life. DG, although uncommon in children, is seen more often in nonwhite admissions.

Gastroparesis (Gp) is a chronic disorder defined by delayed gastric emptying in the absence of an obstructing structural lesion.¹ It commonly presents with symptoms such as nausea, vomiting, early satiety, postprandial fullness, and in some patients, abdominal pain.

Gp often can be associated with severe symptoms, for which treatment may require hospitalization.² There has been an increase in emergency department visits^3–5 and hospitalization of patients with the diagnosis of Gp, leading to an increase in hospital expenditures. The emergency department visits and hospitalizations are likely due to the difficulty in the management of symptoms, dehydration, nutrition, as well as medical procedures.^5–7 Longitudinal studies suggest that 60% to 75% of patients with Gp require hospitalization at some time.^7,8 Prolonged hospitalizations, as well as patients that are frequently readmitted, account for a sizable proportion of hospital utilization and costs.⁶

The true prevalence of Gp is not known.⁹ There have been few epidemiological studies on the population prevalence of the disease.⁵ One epidemiological study, conducted in Olmsted County in Minnesota, showed a female predominance with a prevalence of 9.6 per 100,000 men and 37.8 per 100,000 women.⁷ There is even more scarcity of data on the incidence and prevalence of Gp in children.^10,11

In adults, Gp is more common in female patients^7,12 but gender distribution is not well described in pediatric patients with Gp. Lu et al¹³ in their review of a nationwide database, did not note a significant gender difference among children hospitalized. The 2 most common forms of Gp are idiopathic and diabetes in adult patients.^7,12 It is estimated that 25% to 40% of adult patients with type 1 diabetes have Gp.¹⁴ Adult female patients are more likely to have idiopathic gastroparesis (IG), and among races, Hispanic patients are more likely to have diabetic gastroparesis (DG).¹⁵ Waseem et al,¹⁶ in a retrospective study, noted that IG to be the most common form of Gp in children. The age at which diabetes becomes a significant contributing factor in Gp is unknown.

In this study, we sought to evaluate gender, race, and etiology distribution for the different age groups of patients admitted with Gp. We did this by examining patient demographics of those who were hospitalized using 2 nationwide databases: Nationwide Inpatient Sample (NIS) and Kid’s Inpatient Database (KID). We also examined the gender and race differences between patients with DG and non-DG admissions.

MATERIALS AND METHODS

Databases

The NIS is a publicly available databases of all-payer hospital inpatient stays that can be used to generate national estimates of admissions. This database is a part of the HCUP (Healthcare Cost and Utilization Project) that is sponsored by the Agency for Healthcare Research and Quality (AHRQ) in Rockville, MD. NIS database was queried for the year 2016.

The NIS is sampled from the State Inpatient Databases (SID), which include all inpatient data that are currently contributed to HCUP. The 2016 NIS sampling frame includes data from 47 statewide data organizations (46 states plus the District of Columbia), covering >97% of the US population and including almost 96% of discharges from US community hospitals.

For comparison, KID, which is the largest publicly available all-payer pediatric inpatient care database in the United States (also a part of AHRQ), was also examined. The KID database includes a somewhat larger pediatric inpatient sample size from 4200 US community hospitals with variable sample rates (10% for newborns and 80% for complicated newborn and other pediatric discharges) compared with NID which enables analyses of rare conditions (eg, Gp).

For discharge weighting in both databases, hospitals are poststratified on 6 characteristics contained in the AHA hospital files—ownership/control, bed size, teaching status, rural/urban location, and US region, with the addition of a stratum for freestanding children’s hospitals in KID database. The large sample size of these databases enables analyses of rare conditions, uncommon treatments, and special patient populations. Both databases contain up to 30 diagnostic codes (Dx1-Dx30) for each admission. Primary diagnosis (Dx1) is the principal diagnosis defined as the condition established to be chiefly responsible for the admission of the patient to the hospital for care.

Data Procurement

Data were examined for the calendar year 2016. A Gp admission was defined as a record with an International Classification of Diseases (ICD)-10 code K318.4 for Gp as the primary diagnosis (Dx1) or as the secondary diagnosis (Dx2) where the first diagnosis was a Gp-related symptom including nausea, vomiting, abdominal pain, dehydration, malnutrition, or weight loss (ICD-10 codes of R11.0, R11, R11.1, R11.11, R11.12, R11.10, R11.14, R11.2, R12, R14, R14.3, R10.10, R10.8, R10.84, R10, R10.9, R10.1, R14.1, R68.81, E46, E64, E44.1, R63.4, R64.1, E86.0). Patients with diabetes were identified using all variants of ICD-10 codes: E08.xx (Diabetes mellitus due to underlying condition), E09.xx (Drug or chemical induced diabetes mellitus), E10.xx (Type 1 diabetes mellitus), E11.xx (Type 2 diabetes mellitus), E13.xx (Other specified diabetes mellitus) among Dx1-Dx30 codes. Similarly, ICD-10 codes were used to identify symptoms and conditions outlined in the Supplemental Tables (Supplemental Digital Content 1, https://links.lww.com/JCG/A773).

All admissions between January 1 and December 30, 2016, were included. Transfers from other hospitals during the same admission were excluded to avoid duplication.

Statistical Analysis

Appropriate weighting factors included in the database provided by AHRQ, based on the hospital poststratification method mentioned above, were applied to the database to ensure they are representative of US inpatient admissions in the year 2016 and are comparable to each other. Patients were categorized in 8 common age groups: infants (ages 0 to 1), preschool (ages 2 to 5), child (ages 6 to 12), teen/adolescent (ages 13 to 20), young adults (ages 21 to 39), middle age (ages 40 to 64), senior (ages 65 to 79), and elderly (ages 80 to 90). Results are expressed as a percent or mean±SD. The χ² and Fisher exact tests were used to examine the difference of categorical variables between groups. The Mantel-Haenszel test was used to analyze trends. Results were significant at P-value <0.05. Post hoc analyses were performed using adjusted standardized residuals with Bonferroni correction to detect between-group differences. All analyses were performed with commercially available SPSS software, version 23. The data detail the admissions percentage out of the total Gp admissions in each age group. We also report the admissions as per 100,000 admissions to better enable comparisons between age groups (and other groups too).

RESULTS

Demographics

The demographics of Gp admissions are shown in Table 1. After properly weighting as described above, there were a total of 15,790 admissions for Gp during 2016 for all age groups from 0 to 90 years which included 90 (0.6%) in the infant group, 100 (0.6%) in preschool groups, 215 (1.4%) in child group, 645 (4.1%) in teen and adolescent group, 4955 (31.4%) in the young adult group, 7250 (45.9%) in middle-age group, 1905 (12.1%) in the senior group, and 630 (4%) in the elderly group. The mean age was 46.1±18.0 years among Gp admissions. Overall, Gp admissions accounted for 0.04% of the 35,668,311 admissions in the NIS database. Rates per 100,000 total admissions are shown in Table 1, which show a similar progressive increase in Gp admissions across different age groups until the middle-age group.

TABLE 1 - Demographics for Hospital Admissions With Gastroparesis

	N (/100,000 Total Admissions)
Admission Characteristics	Total	Infants (0-1)	Preschool (2-5)	Child (6-12)	Teen/Adolescent (13-20)	Young Adults (21-39)	Middle Age (40-64)	Senior (65-79)	Elderly (80-90)
Total admissions	32,799,061	4,196,461	247,920	322,900	961,704	6,448,299	9,191,873	7,076,749	4,353,156
Gastroparesis admissions	15,790 (48.14)	90 (2.14)	100 (40.33)	215 (66.58)	645 (67.06)	4955 (76.84)	7250 (78.87)	1905 (26.91)	630 (14.47)
Age (mean±SD)	46.15±18.0	0.28±0.45	2.95±0.93	9.23±2.05	16.70±2.28	31.53±5.20	51.18±7.04	70.93±4.14	84.67±3.27
Gender
Female	11,945 (36.41)	55 (1.31)	45 (18.15)	135 (41.8)	495 (51.47)	3905 (60.55)	5445 (59.23)	1395 (19.71)	470 (10.79)
Male	3835 (11.69)	35 (0.83)	55 (22.18)	80 (24.77)	150 (15.59)	1050 (16.28)	1795 (19.52)	505 (7.13)	160 (3.67)
Diabetic gastroparesis
Total	3995 (12.18)	0 (0)	0 (0)	5 (1.54)	40 (4.15)	990 (15.35)	2255 (24.53)	535 (7.55)	170 (3.9)
Female	2715 (8.27)	0 (0)	0 (0)	5 (1.54)	35 (3.63)	665 (10.31)	1490 (16.2)	400 (5.65)	120 (2.75)
Male	1280 (3.9)	0 (0)	0 (0)	0 (0)	5 (0.51)	325 (5.04)	765 (8.32)	135 (1.9)	50 (1.14)
Race
White	9980 (30.42)	45 (1.07)	40 (16.13)	115 (35.61)	420 (43.67)	2900 (44.97)	4595 (49.98)	1375 (19.42)	485 (11.14)
African American	3255 (9.92)	25 (0.59)	15 (6.05)	35 (10.83)	60 (6.23)	1195 (18.53)	1615 (17.56)	265 (3.74)	45 (1.03)
Hispanic	1370 (4.17)	5 (0.11)	20 (8.06)	40 (12.38)	95 (9.87)	485 (7.52)	570 (6.2)	115 (1.62)	40 (0.91)
Asian/Pacific Islander	245 (0.74)	0 (0)	0 (0)	10 (3.09)	10 (1.03)	75 (1.16)	80 (0.87)	40 (0.56)	30 (0.68)
Native American	105 (0.32)	0 (0)	0 (0)	0 (0)	0 (0)	40 (0.62)	45 (0.48)	20 (0.28)	0 (0)
Other	320 (0.97)	5 (0.11)	10 (4.03)	10 (3.09)	25 (2.59)	90 (1.39)	125 (1.35)	35 (0.49)	20 (0.45)
Median household income
$1-$42,000	5335 (16.26)	45 (1.07)	15 (6.05)	50 (15.48)	165 (17.15)	1790 (27.75)	2480 (26.98)	590 (8.33)	200 (4.59)
$43,000-$53,999	4270 (13.01)	10 (0.23)	10 (4.03)	50 (15.48)	170 (17.67)	1245 (19.3)	2025 (22.03)	580 (8.19)	180 (4.13)
$54,000-$70,999	3500 (10.67)	30 (0.71)	45 (18.15)	60 (18.58)	125 (12.99)	1090 (16.9)	1600 (17.4)	410 (5.79)	135 (3.1)
>$71,000	2500 (7.62)	5 (0.11)	30 (12.1)	55 (17.03)	175 (18.19)	780 (12.09)	1060 (11.53)	290 (4.09)	105 (2.41)
Insurance
Medicare	5890 (17.95)	0 (0)	0 (0)	0 (0)	5 (0.51)	1065 (16.51)	2530 (27.52)	1700 (24.02)	590 (13.55)
Medicaid	3720 (11.34)	70 (1.66)	55 (22.18)	85 (26.32)	245 (25.47)	1710 (26.51)	1520 (16.53)	25 (0.35)	5 (0.11)
Private insurance	5035 (15.35)	20 (0.47)	40 (16.13)	130 (40.26)	350 (36.39)	1685 (26.13)	2625 (28.55)	160 (2.26)	25 (0.57)
Self-pay	740 (2.25)	0 (0)	0 (0)	0 (0)	20 (2.07)	355 (5.5)	355 (3.86)	5 (0.07)	5 (0.11)
No charge	60 (0.18)	0 (0)	0 (0)	0 (0)	0 (0)	35 (0.54)	25 (0.27)	0 (0)	0 (0)
Other	345 (1.05)	0 (0)	5 (2.01)	0 (0)	25 (2.59)	105 (1.62)	190 (2.06)	15 (0.21)	5 (0.11)
Hospital metrics
Total hospital cost (mean±SD) ($)	40,539.42±62487.01	74,664.46±136232.07	41,567.01±38977.89	31,237.16±20704.97	42,397.69±44066.81	36,474.52±39189.84	40,765.13±66313.01	42,377.51±39673.01	60,530.73±156,753.92
Hospital stay (mean±SD) (d)	4.69±4.97	8.44±8.00	6.90±5.92	4.35±3.27	4.98±5.34	4.45±5.16	4.52±4.44	5.09±5.18	6.43±6.89
Elective admission	1840 (5.6)	10 (0.23)	10 (4.03)	40 (12.38)	150 (15.59)	510 (7.9)	830 (9.02)	255 (3.6)	35 (0.8)
Urgent admission	13,940 (42.5)	80 (1.9)	90 (36.3)	175 (54.19)	495 (51.47)	4440 (68.85)	6410 (69.73)	1650 (23.31)	595 (13.66)
Disposition
Died during admission	20 (0.06)	0 (0)	0 (0)	0 (0)	0 (0)	0 (0)	5 (0.05)	5 (0.07)	10 (0.22)
Home or self-care	12,440 (37.92)	80 (1.9)	95 (38.31)	210 (65.03)	565 (58.74)	4170 (64.66)	5820 (63.31)	1250 (17.66)	250 (5.74)
Transfer to short-term hospital	185 (0.56)	5 (0.11)	0 (0)	0 (0)	15 (1.55)	90 (1.39)	55 (0.59)	20 (0.28)	0 (0)
Transfer to SNF, ICF	715 (2.17)	0 (0)	0 (0)	0 (0)	0 (0)	40 (0.62)	235 (2.55)	235 (3.32)	205 (4.7)
Home health care	2035 (6.2)	5 (0.11)	5 (2.01)	5 (1.54)	65 (6.75)	435 (6.74)	965 (10.49)	385 (5.44)	165 (3.79)
Left against medical advice	385 (1.17)	0 (0)	0 (0)	0 (0)	0 (0)	215 (3.33)	160 (1.74)	10 (0.14)	0 (0)

There were 10 admissions with missing genders and 515 admissions had missing race.

ICF indicates intermediate care facility; SNF, skilled nursing facility.

Gender Distribution of Gp Admissions

There were a total of 11,945 (75.6%) female and 3835 (24.3%) male admissions for Gp. The infant group consisted of 55 (61.1%) female admissions and 35 (38.9%) male admissions. There were 45 (45%) female admissions in the preschool group, 135 (62.8%) in the child group, and 495 (76.7%) in the teen/adolescent group. Female admission predominance continued throughout adulthood with 3905 (78.8%) in young adults, 5445 (75.1%) in the middle-age group, 1395 (73.2%) in seniors, and 470 (74.6%) in elderly. When examining gender differences at each specific age in children, female admissions showed a distinct relative increase at 12 years of age in both NIS and KID databases (Fig. 1). Rates per 100,000 total admissions are shown in Table 1, which again shows a similar trend with more female admissions, especially after the second decade, whereas Gp admission rates for male patients, although statistically significant, showed a less dramatic increase (all P<0.05).

FIGURE 1: Gender Distribution of gastroparesis admissions in NIS and KID 2016 Databases for Children (ages 0 to 20). Both NIS (A) and KID (B) databases show an increase in female admissions at age 12. CI indicates confidence interval; KID, Kid’s Inpatient Database; NIS, Nationwide Inpatient Sample.

Etiology of Gp

NIS reported a total of 3995 admissions of patients with DG (25.3% of all admissions), of which 2715 (68%) were female, and 1280 (32%) were male. There were only 45 DG admissions under age 20, with 5 (0.1%) admissions in the child group and 40 (1%) among teens/adolescents. The comparison between DG and non-DG admissions in NIS and KID databases are shown in Figure 2.

FIGURE 2: Diabetes among gastroparesis admissions in Nationwide Inpatient Sample (NIS) (A) and Kid’s Inpatient Database (KID) (B) 2016 Databases.

Despite the fact that most DG admissions were female, a similar percentage of diabetics in male patients with Gp and female patients with Gp admissions were seen in most groups. The exceptions were young adults [male: 325 (31.0% of male Gp admissions), female: 665 (17.0% of female admissions), P<0.01] and middle-age group [male: 765 (42.6%), female: 1490 (27.4%), P<0.01] where male admissions with Gp had a higher percentage of diabetes compared with female admissions with Gp (Table 2).

TABLE 2 - Comparison of Diabetic and Nondiabetic Gastroparesis Admissions

	n (%)
Admission Characteristics	Total	Infants (0-1)	Preschool (2-5)	Child (6-12)	Teen/Adolescent (13-20)	Young Adults (21-39)	Middle Age (40-64)	Senior (65-79)	Elderly (80-90)
Gender
Female
Nondiabetic	9230 (77.28)	550 (100)	45 (100)	130 (96.3)	460 (92.93)	3240 (82.98)	3955 (72.64)	995 (71.33)	350 (74.47)
Diabetic	2715 (22.73)	0 (0)	0 (0)	5 (3.71)	35 (7.08)	665 (17.03)	1490 (27.37)	400 (28.68)	120 (25.54)
Total	11,945	55	45	135	495	3905	5445	1395	470
Male
Nondiabetic	2555 (66.63)	35 (100)	55 (100)	80 (100)	145 (96.67)	725 (69.05)	1030 (57.39)	370 (73.27)	110 (68.75)
Diabetic	1280 (33.38)	0 (0)	0 (0)	0 (0)	5 (3.34)	325 (30.96)	765 (42.62)	135 (26.74)	50 (31.25)
Total	3835	35	55	80	150	1050	1795	505	160
P	0.000*	—	—	0.436	0.680	0.000*	0.000*	0.797	0.645
Race
White
Nondiabetic	8205 (82.21)	45 (100)	40 (100)	115 (100)	400 (95.3)	2595 (89.48)	3535 (76.93)	1085 (78.9)	385 (79.38)
Diabetic	1775† (17.78)	(0)	(0)	0 (0)	20 (4.8)	305† (10.51)	1060† (23.06)	290† (21.09)	100† (20.61)
Total	9980	45	40	115	420	2900	4595	1375	485
African American
Nondiabetic	2010 (61.75)	25 (100)	15 (100)	30 (85.71)	55 (91.7)	780 (65.27)	925 (57.27)	150 (56.6)	30 (66.66)
Diabetic	1245† (38.24)	(0)	(0)	5† (14.28)	5 (8.4)	415† (34.72)	690† (42.72)	115† (43.39)	15 (33.33)
Total	3255	25	15	35	60	1195	1615	265	45
Hispanic
Nondiabetic	835 (60.94)	5 (100)	20 (100)	40 (100)	85 (89.5)	325 (67.01)	280 (49.12)	70 (60.86)	10 (25)
Diabetic	535† (39.05)	(0)	(0)	0 (0)	10 (10.6)	160† (32.98)	290† (50.87)	45 (39.13)	30† (75)
Total	1370	5	20	40	95	485	570	115	40
Native American
Nondiabetic	40 (38.09)	(0)	(0)	(0)	0 (0)	15 (37.5)	20 (44.44)	5 (25)	(0)
Diabetic	65† (61.9)	(0)	(0)	(0)	0 (0)	25† (62.5)	25 (55.55)	15 (75)	(0)
Total	105	0	0	0	0	40	45	20	0
Asian/Pacific Islander
Nondiabetic	170 (69.38)	(0)	(0)	10 (100)	10 (100)	55 (73.33)	60 (75)	20 (50)	15 (50)
Diabetic	75 (30.61)	(0)	(0)	0 (0)	0 (0)	20 (26.66)	20 (25)	20 (50)	15 (50)
Total	245	0	0	10	10	75	80	40	30
Other
Nondiabetic	220 (68.75)	5 (100)	10 (100)	10 (100)	25 (100)	75 (83.33)	65 (52)	20 (57.14)	10 (50)
Diabetic	100 (31.25)	(0)	(0)	0 (0)	0 (0)	15 (16.66)	60 (48)	15 (42.85)	10 (50)
Total	320	5	10	10	25	90	125	35	20
P	0.000*	—	—	0.275	0.821	0.000*	0.000*	0.001*	0.007*
Median household income
$1-42,999
Nondiabetic	3735 (70)	45 (100)	15 (100)	50 (100)	150 (91)	1290 (72.06)	1635 (65.92)	425 (72.03)	125 (62.5)
Diabetic	1600† (29.99)	0 (0)	0 (0)	0 (0)	15 (9.1)	500† (27.93)	845† (34.07)	165 (27.96)	75 (37.5)
Total	5335	45	15	50	155	1790	2480	590	200
$43,000-$53,999
Nondiabetic	3140 (73.53)	10 (100)	10 (100)	45 (90)	160 (94.2)	1015 (81.52)	1360 (67.16)	405 (69.82)	135 (75)
Diabetic	1130 (26.46)	0 (0)	0 (0)	5 (10)	10 (5.9)	230 (18.47)	665 (32.83)	175 (30.17)	45 (25)
Total	4270	10	10	50	170	1245	2025	580	180
$54,000-$70,999
Nondiabetic	2740 (78.39)	30 (100)	45 (100)	60 (100)	115 (92)	945 (86.69)	1140 (71.25)	285 (69.51)	120 (88.88)
Diabetic	755† (21.6)	0 (0)	0 (0)	0 (0)	10 (8)	145† (13.3)	460 (28.75)	125 (30.48)	15 (11.11)
Total	3495	30	45	60	125	1090	1600	410	135
>$71,000
Nondiabetic	2055 (82.2)	5 (100)	30 (100)	55 (100)	170 (97.2)	690 (88.46)	805 (75.94)	230 (79.31)	70 (66.66)
Diabetic	445† (17.8)	0 (0)	0 (0)	0 (0)	5 (2.9)	90† (11.53)	255† (24.05)	60 (20.68)	35 (33.33)
Total	2500	5	30	55	175	780	1060	290	105
P	0.000*	—	—	0.337	0.736	0.000*	0.039*	0.561	0.106

^*χ² test was used to compare groups. P<0.05 was considered statistically significant.

^†Post hoc tests with Bonferroni correction were done to identify between-group differences. P<0.05 was considered statistically significant.

Ethnicity and Race

The most common race was white, with 9980 admissions (63.2%) overall. By age group, 45 (50%) of the infant Gp admissions were whites, in preschool group 40 (40%), in the child group with 115 (53.5%), in teens/adolescents with 420 (65.1%), in young adults with 2900 (58.5%), in middle-age patients with 4595 (63.4%), in seniors with 1375 (72.2%), and in elderly with 485 (77%). Table 1, for rates per 100,000 total admissions, show the same race distribution meaning white as the most common race, followed by African American and Hispanics among Gp admissions.

The second and third most common races in NIS were African American with 3255 (20.6%) and Hispanics with 1370 (8.7%). The second most common race in infants was African American, with 25 admissions (27.8%). In the preschool, child, and teen/adolescent groups, the second most common race was admissions of Hispanic patients [preschool 20 (20%), child 40 (18.6%), and teen/adolescent 95 (14.7%)]. African American remained the second most common race among Gp admissions throughout adulthood [young adult: 1195 (24.1%), middle age: 1615 (22.3%), senior: 265 (13.9%), elderly: 45 (7.1%)].

Female Gp admissions were more common in all races with 7610 (76.3%) among whites, 2420 (74.3%) among African Americans, 1050 (76.6%) among Hispanics, 155 (63.3%) among Native Americans, 85 (81%) among Asian/Pacific Islanders, and 250 (78.1%) among other races.

No difference between DG and non-DG admissions was observed in infant through adolescent groups among races, but in the young adult group, diabetes was present in 25 (62.5%) of native American Gp admissions, 415 (34.7%) of African American admissions, 160 (33%) of Hispanic, and 305 (10.5%) of white admissions (P<0.01). In middle-age patients, diabetes was present in 290 (50.9%) of Hispanics, 690 (42.7%) of African America, and 1060 (23.1%) of white Gp admissions (P<0.01). In the senior group, diabetes was more common in native Americans with 15 (75%) Gp admissions and African Americans with 115 (43.4%) admissions, but less common among whites with 290 (21.1%) admissions (P<0.01). In the elderly group, diabetes was present in 30 (75%) of Hispanics and 100 (20.6%) of white Gp admissions (P<0.01).

Hospital Metrics

Gp was the first ICD-10 diagnosis code (Dx1) in 15,095 (95.6%) of NIS admissions, and as the second diagnosis code (Dx2) with 700 (4.4%). Among admissions with Gp as Dx1, the top Dx2 codes were N17.9 (acute kidney failure) in 810 (5.4%), E43 (severe protein-calorie malnutrition) in 785 (5.2%), and I 10 (essential hypertension) in 750 (5.0%) of these admissions.

Among admissions with Gp as Dx2, the top 3 Dx1 in NIS, were R11.2 (unspecified nausea and vomiting) in 245 (35%), R10.9 (unspecified abdominal pain) in 215 (30.7%), E86.0 (dehydration) in 150 (21.4%) admissions.

Mean hospital length of stay was 4.7±5.0 days (infant: 8.4±8.0, preschool: 6.9±5.9, child: 4.3±3.3, teen: 4.9±6.1, adolescent: 5.1±4.1, young adults: 4.4±5.2, middle age: 4.5±4.4, senior: 5.1±5.2, elderly: 6.4±6.9).

The majority of the admissions were nonelective [13,940 (88.3%)]. This pattern was observed across different age groups. Twenty (0.1% of Gp admissions) of patients died during the hospital admission (5 in the middle-age group, 5 in the senior groups, and 10 in the elderly group), 12,440 (78.8%) were discharged home, 2035 (12.9%) were discharged with home health care, 715 (4.5%) were discharged to a skilled nursing facility, 185 (1.2%) were transferred to a short-term hospital, and 385 (2.4%) left against medical advice.

Nutrition and Weight

Malnutrition (ICD-10 codes: E40-E46) was present in 2280 (14.4%) of admissions in NIS database [infant: 30 (33.3%), preschool: 10 (10%), child: 15 (7%), teen/adolescent: 85 (13.2%), young adult: 660 (13.3%), middle age: 945 (13%), senior: 375 (19.7%), elderly: 160 (25.4%)] (Supplemental Table 1, Supplemental Digital Content 1, https://links.lww.com/JCG/A773).

Abnormal weight loss (ICD-10 code: R63.4) was seen in 620 (3.9%) admissions were again most common in the teen/adolescent group with 60 (9.4%). Underweight status was reported for 75 (0.5%) of admissions.

Obesity (ICD-10 code: E66) was reported in 2035 (12.9%) Gp admissions. Vitamin B₁₂ deficiency was present in 15 (0.1%) Gp admissions, and vitamin D deficiency was present in 255 (1.6%) of Gp admissions.

Alcohol and Substance Use

Smoking was reported in 3075 (19.5%) Gp admissions [teen/adolescent: 30 (4.7%), young adults: 1235 (24.9%), middle age: 1620 (22.3%), senior: 185 (9.7%), elderly: 5 (0.8%)] (Supplemental Table 2, Supplemental Digital Content 1, https://links.lww.com/JCG/A773). Alcohol use was seen in 300 (2.1%) Gp admissions [teen/adolescent: 10 (1.6%), young adults: 80 (1.6%), middle age: 180 (2.5%), senior: 50 (2.6%), elderly: 10 (1.6%)]. Cannabinoid use was present in 1165 (7.4%) Gp admissions [teen/adolescent: 70 (10.9%), middle age: 650 (13.1%), senior: 435 (6%), elderly: 10 (0.5%)]. Opioid use was reported in 1095 (6.9%) Gp admissions [young adults: 400 (8.1%), middle age: 615 (8.5%), senior: 80 (4.2%)].

Analysis of KID Database

The results of the KID database assessing pediatric population primarily complemented the NIS database, which assessed all ages (Supplemental Tables 3, 4, Supplemental Digital Content 1, https://links.lww.com/JCG/A773). There were 6,263,398 weighted total admissions in the KID database of the year 2016 in contrast to 5,726,304 admissions of children of 20 y and younger in the NIS database), of which there were 1114 Gp admissions (67.7% female) with an average age of 12.6±6.6 years of children (ages 0 to 20 y) were present in KID database of the year 2016 (see Supplemental Table 3, Supplemental Digital Content 1, https://links.lww.com/JCG/A773, for per capita rates). Only 2.9% of Gp admissions had DG, mostly in the teen/adolescent group. White was the most common race (63.2%), followed by Hispanic (13.9%) and African American (11.8%). No gender difference (P=0.121) or race difference (P=0.198) was seen between DG and non-DG admissions.

DISCUSSION

In this retrospective epidemiological study, we used nationwide databases to examine the demographic features in patients of different age groups who were hospitalized due to Gp. Two important findings in this large database study are related to the gender of patients and the etiology of Gp. Although there is a female predominance in the Gp hospitalizations, this was age-related, with a significant increase in female admissions in the second decade of life. DG was an important etiologic factor in patients with Gp, but this was also age-related and gender-related. DG was uncommon in children but became increasingly an important etiology in the older age groups. These findings were also present when comparisons were made with admissions expressed as per total admissions or per 100,000 admissions.

In general, hospitalization for Gp is on the rise both in adults^2,6 and children.¹³ The age distribution of children with Gp is not well defined. Lu et al¹³ reviewed a national database between the years 2004 and 2013 and noticed higher admissions in the first 5 years of life among children. In our review of the NIS database, the total number of admissions for Gp in children, albeit less common than adults in general, increased with age, and young adults and middle-aged patients had the highest admissions overall.

In adults, Gp is generally more common in female patients.⁷ Studies from the National Institutes of Health (NIH) Gastroparesis Consortium showed a female predominance in IP in adult patients¹² and in another study in both idiopathic and diabetic forms.¹⁷ Lu et al,¹³ in their review of nationwide data, noted no significant gender difference in children. In our study looking at Gp across the ages, the rates of female admissions started to increase in NIS (as well as KID) databases at the age of 12 years. Based upon data from the Third National Health and Nutrition Examination Survey (NHANES III) database, the median age of menarche in the United States is 12.4 years,¹⁸ this increase in Gp at age 12 could be explained by hormonal involvement in the pathogenesis of Gp in young girls after puberty. Progesterone has been shown to delay gastric emptying, and estrogen has been shown to control nitric oxide-mediated gastric emptying in mice which further supports the role that hormones may play in Gp.¹⁹ In humans, postmenopausal women have been shown to have faster gastric emptying than premenopausal women.²⁰

Most cases of Gp in adults are idiopathic with diabetes and other etiologies such as the postsurgical, viral, medication-related following in frequency.^8,13 Observational studies have found diabetes to be a minor etiology of Gp in children.^16,21 This may be due to the need for time for diabetic complications to develop. In the NIH Gastroparesis Consortium study on diabetic complications in DG, the duration from diabetes diagnosis to the onset of Gp symptoms was twelve years.²² When examining the NIS database, the prevalence of diabetic children admitted for Gp was low but increased in the second decade of life. Although per capita rate of DG was higher in females across all age groups, there was a greater percentage of diabetic admissions in young adults and middle-aged patients in male patients compared with females, where for these age groups, there was a predominance of non-DG. It is unclear whether this is due to the higher prevalence of diabetes in male patients of these age groups, higher acuity of symptoms in this group leading to more admissions, or perhaps a higher rate of IG among female admissions. Interestingly, the NIH Gastroparesis Consortium studies suggest a higher proportion of women had Gp of idiopathic etiology (69%) than men (46%); women had more severe symptoms of stomach fullness, early satiety, postprandial fullness, bloating, stomach visibly larger, and upper abdominal pain.¹⁵ A review of the Nationwide Readmission Database (NRD) for the year 2014 of 5268 adult patients admitted with Gp showed higher admission rates in younger adults but also a lower readmission rate for patients with diabetes.²³

The most common race in the majority of admissions in both databases across different age groups was white. The next most common races were African American and Hispanic. These data compare to the NIH Gastroparesis registry studies of outpatients with Gp, where whites comprised 77% of patients, Hispanic ethnicity 12%, and African Americans 9%.¹⁵ The rate of admissions in diabetic children was not different among races, but in all adult age groups in NIS, Native American, African American, and Hispanic admissions showed a higher percentage of Gp patients with diabetes. Again, it is not clear whether this is due to the higher prevalence of diabetes in these groups or higher admission rates. A large prospective study of 718 adult patients in the NIH Gastroparesis Consortium database showed a higher rate of Gp among white patients compared with other races (77%) with overall female predominance (84%) and a significantly higher proportion of Hispanic patients with DG (59%).¹⁵ They also showed a lower rate of admissions among women compared with men.

The evolution of Gp from children to adults is poorly understood, and our study used a large national database to examine different trends in Gp patient admissions of different ages. This database provides a large sample of admissions which is representative of US population hospital admissions. This study, however, is not without limitations. NIS database does not contain unique patient information, which would be ideal to study prevalence. This study relies on accurate coding during these admissions. The use of administrative codes for Gp may not be a reliable indicator of the presence or absence of disease. There is a chance of misclassification or administrative errors. Gp is strictly defined by the presence of delayed gastric emptying in the absence of obstruction. Clinical suspicion without a gastric emptying study may be used in some tertiary centers; this had been estimated to possibly account for up to 40% of cases.⁸ As with any analysis using a database of this type, there are data not included in the collected variables recorded that would be useful to assess. Some patient populations, such as the elderly are more prone to higher admission rates in general due to other comorbidities, and this may confound the true admission rates in those populations. Some of these variables include race or ethnicity, hemoglobin A1c level, gastric emptying study results, home medication use, and prior therapies for Gp. Finally, this database looks at patients admitted for Gp and not the individuals with Gp in the community.

In summary, this study used the NIS national admission database, which provides data on large cohorts of hospitalized patients allowing a unique opportunity to study the characteristics of less common conditions such as Gp. Although there is a female predominance in Gp admissions, this trend increases remarkably during teenage years and rising throughout the adult years. Diabetes is an important comorbidity in Gp in adults. DG was not commonly seen in children; it started to increase in the second decade of life. DG was also more common in male and nonwhite young and middle-aged adults. We also looked at a pediatric database (KID) which supported the findings from the NIS database in adults showing a lower percentage of female Gp patients in the younger patients as well as low numbers of DG in young patients. This study illustrates that comprehensive epidemiological studies of hospitalization of patients of various ages with Gp can help us understand factors associated with this condition.