Vol. 117 No. 4 April 2014

Characteristics of hospitalizations attributed to herpeticgingivostomatitis: analysis of nationwide inpatient sampleVeerasathpurush Allareddy, BDS, MBA, MHA, PhD, MMSc,a and Satheesh Elangovan, BDS, DSc, DMScb

University of Iowa, Iowa City, IA, USA

Objective. Herpetic gingivostomatitis (HGS) is a prevalent oral condition of viral origin. Some patients with HGS visit

emergency departments for treatment. This study is aimed at determining the nationwide estimates of hospitalizations due to

HGS in the United States.

Study Design. The Nationwide Inpatient Sample of the Agency for Healthcare Research and Quality was used for the years

2004 to 2010. Hospital admissions with a primary diagnosis of HGS were identified using International Classification of

Diseases (ICD)-9-CM code 054.2.

Results. During the study period, a total of 12 536 hospitalizations were attributed to HGS, and the number of hospitalizations

per year (mean, 1791) remained fairly constant during the years examined. The median age was 3.2 years. The mean age (15.6

years for the overall period) ranged from a low of 15.2 in 2010 to a high of 16.8 in 2007. The total charges that resulted from

these hospitalizations were $229.4 million.

Conclusions. This study further confirms that substantial resources are spent in treating oral conditions in hospital settings.

(Oral Surg Oral Med Oral Pathol Oral Radiol 2014;117:471-476)

The prevalence of orofacial infections caused by herpessimplex virus (HSV) in humans is widespreadthroughout the world.1,2 Oral, facial, and ocular in-fections are caused by HSV type 1, whereas HSV type2 typically causes genital lesions but is also implicatedin some cases of oral infections.3 The primary form ofHSV-1 oral infection is herpetic gingivostomatitis(HGS), which usually affects patients aged 6 months to5 years or patients in their early twenties.4 Patients withHGS usually present with pain, fever, lymphadenopa-thy, malaise, irritability, and headache. HSV-1 has alsobeen implicated in viral encephalitis, lymphadenitis,and severe ocular lesions.5 Only a small percentage(about 1%) of patients with HGS exhibit clinical man-ifestations, despite widespread exposure to this viralagent.6

Several studies in the past have investigated theseroprevalence of HSV-1 in various populations, but toour knowledge, little is known about the prevalence ofhospitalizations of patients diagnosed primarily withHGS.7 In our previous study, using the NationwideEmergency Department Sample (NEDS), which is oneof the databases of the Healthcare Cost and UtilizationProject (HCUP) sponsored by the Agency for Health-care Research and Quality (AHRQ), we identified that

aAssociate Professor, Department of Orthodontics, University of IowaCollege of Dentistry.bAssistant Professor, Department of Periodontics, University of IowaCollege of Dentistry.Received for publication Nov 26, 2013; returned for revision Jan 14,2014; accepted for publication Jan 20, 2014.� 2014 Elsevier Inc. All rights reserved.2212-4403/$ - see front matterhttp://dx.doi.org/10.1016/j.oooo.2014.01.022

close to 23 124 emergency department (ED) visits wereattributed to HGS in 2007.8 Of these visits, close to69.5% occurred in those aged 19 years or younger, andfemale patients accounted for 55.7% of all ED visits.8

In line with our previous investigation, the objectives ofthis assessment are 2-fold: (1) to provide nationwideestimates of hospitalizations in the United States due toHGS using the Nationwide Inpatient Sample (NIS)(another HCUP-AHRQ database)9 for the years 2004 to2010 and (2) to examine the sociodemographic char-acteristics and hospital-related outcomes associatedwith these hospitalizations.

MATERIALS AND METHODSDatabaseThe NIS for 2004 to 2010 was used for the presentstudy. The NIS is a 20% stratified probability sample ofall acute-care hospitals in the United States. The sampleis drawn based on stratification factors including hos-pital size (by bed count), location, region, ownership,and teaching status.9 Each hospital selected into thesample provides information on all hospitalizationsoccurring in the selected years. Each hospitalization inthe NIS database is assigned a sampling weight, whichcan be used to project all estimates and outcomes to

Statement of Clinical Relevance

This study found that a substantial number of pa-tients with herpetic gingivostomatitis are hospital-ized in the United States each year, and it identifiedhigh-risk cohorts who are likely to be hospitalized.

471

Table I. Characteristics of hospitalizations due toherpetic gingivostomatitis (2004 to 2010)

Characteristic Response N (%)

Year of hospitalization 2004 1933 (15.4)

ORAL MEDICINE OOOO

472 Allareddy and Elangovan April 2014

nationally representative levels of all hospitalizationsoccurring in the United States annually. The NIS da-tabases are publically available for purchase fromHCUP-AHRQ.

2005 1847 (14.7)2006 1677 (13.4)2007 1629 (13)2008 1952 (15.6)2009 1647 (13.1)2010 1851 (14.8)

Sex Male 5767 (46.4)Female 6651 (53.6)

Primary payer Medicare 1105 (8.8)Medicaid 5645 (45.1)Private insurance 4691 (37.5)Uninsured 727 (5.8)Other insurance plans 342 (2.7)

Race White 4949 (51)

Institutional review board approvalOne of the authors completed the data user agree-ment with HCUP-AHRQ and obtained the NISdatabases. According to the data user agreement,individual cell counts �10 cannot be presented topreclude us from identifying patients uniquely. Inaccordance with this data user agreement, such lowcell counts are not presented in the present study.The present study was exempt from institutional re-view board approval.

Black 1316 (13.6)Hispanic 2522 (26)Asian/Pacific Islander 371 (3.8)Native Americans 32 (0.3)Other races 514 (5.3)

Type of admission Emergency/urgent 10661 (85.4)Elective 1823 (14.6)

Age distribution <1 year (infants) 2699 (21.6)1 to 3 years (toddlers) 3484 (27.9)4 to 5 years (preschool) 647 (5.2)6 to 10 years (school age) 1276 (10.3)11 to 14 years (early

adolescents)490 (3.9)

15 to 17 years (middleadolescents)

550 (4.4)

18 to 21 years (lateadolescents)

489 (3.9)

22 to 30 years 578 (4.6)31 to 40 years 428 (3.4)41 to 50 years 390 (3.1)>50 years 1451 (11.6)

Age (y) Mean 15.6Standard error 0.62

Individual cell counts will not add up to the global totals because ofmissing values.

Selection of hospitalizations and outcomesThe NIS database has 25 diagnoses fields (15 fromyears 2004 to 2008 and 25 from 2009 onward). Ac-cording to the NIS database documentation, the pri-mary diagnosis field presents information on thereason for hospitalization. For the present study, allhospitalizations with a primary diagnosis of HGSwere selected for analysis. Selection was based onInternational Classification of Diseases (ICD)-9-CMcode 054.2. The patient-related characteristics thatwere examined included age distribution, sex, race,disposition status, comorbid burden (as estimatedfrom the secondary diagnoses fields using the NISseverity index files),9 insurance status, hospitalizationcharges, length of stay, and hospital characteristics(including hospital size [by bed count], geographicregion, and teaching status). For age distribution, theAmerican Academy of Pediatrics guidelines were usedto stratify those aged �21 years into groups includinginfants (<1 year), toddlers (1 to 3 years), preschool(4 to 5 years), school age (6 to 10 years), early ado-lescents (11 to 14 years), middle adolescents (15 to 17years), and late adolescents (18 to 21 years). For thoseaged >21 years, the age was grouped as 22 to 30years, 31 to 40 years, 41 to 50 years, and >50 years.Costs were reported in US dollars, and all hospitali-zation charges were inflation-adjusted to 2010 dollarsusing the Bureau of Labor Statistics hospital inpatientcare inflation rates.

Each individual hospitalization was the unit ofanalysis. For projecting all estimates to nationallyrepresentative levels, the discharge weight variableassigned to each hospitalization was used. The complexstratification sample design was accounted for, and theNIS stratum variable was used as the stratification unit.SAS version 9.3 (SAS Institute Inc, Cary, NC, USA)was used to provide descriptive statistics.

RESULTSDuring the study period (2004 to 2010), a total of12 536 hospitalizations were primarily due to HGS. Thenumber of hospitalizations occurring each year wasconsistent, ranging from a low of 1629 in 2006 to ahigh of 1952 in 2007. The characteristics of hospitali-zations are summarized in Table I. The mean age was15.6 years (median, 3.2 years). The mean age rangedfrom a low of 15.2 in 2010 to a high of 16.8 in 2007.Infants (aged <1 year) and toddlers (aged 1 to 3 years)accounted for 21.6% and 27.9% of all hospitalizations,respectively. Female patients comprised 53.6% of allhospitalizations. The primary payers included Medicare(8.8%), Medicaid (45.1%), private insurance plans(37.5%), and other insurance plans (2.7%). A total of

Table II. Presence of comorbid conditions amonghospitalizations due to herpetic gingivostomatitis

Comorbid conditions N (%)

AIDS 0Alcohol abuse 70 (0.6)Deficiency anemias 749 (6)Rheumatoid arthritis/collagen vascular diseases 232 (1.8)Chronic blood loss anemia 24 (0.2)Congestive heart failure 148 (1.2)Chronic pulmonary disease 1127 (9)Coagulopathy 280 (2.2)Depression 360 (2.9)Diabetes, uncomplicated 299 (2.4)Diabetes with chronic complications 78 (0.6)Drug abuse 95 (0.8)Hypertension 1050 (8.4)Hypothyroidism 237 (1.9)Liver disease 215 (1.7)Lymphoma 254 (2)Fluid and electrolyte disorders 8145 (65)Metastatic cancer 209 (1.7)Other neurologic disorders 472 (3.8)Obesity 146 (1.2)Paralysis 138 (1.1)Peripheral vascular disorders 36 (0.3)Psychoses 99 (0.8)Pulmonary circulatory disorders 39 (0.3)Renal failure 294 (2.3)Solid tumor without metastasis 228 (1.8)Peptic ulcer disease: excluding bleeding 0Valvular disease 84 (0.7)Weight loss 517 (4.1)No. of aforementioned comorbid conditions

0 2694 (21.5)1 6681 (53.3)2 1760 (14)3 699 (5.6)4 392 (3.1)5 171 (1.4)6 84 (0.7)7 39 (0.3)�8 15 (0.1)

Table III. Disposition status after hospitalization dueto herpetic gingivostomatitis

Characteristic Response N (%)

Disposition afterhospitalization

Routine 11707 (93.4)Transfer to another

short-term hospital92 (0.7)

Transfer to otherlong-term carefacilities*

289 (2.3)

Home health care 347 (2.8)Against medical advice 33 (0.3)Died in hospital 68 (0.5)

*Includes skilled nursing facility, intermediate care facility, or anothertype of facility.

Table IV. Hospital characteristics

Characteristic Response N (%)

Hospital region Northeast 2715 (21.7)Midwest 2479 (19.8)South 5181 (41.3)West 2161 (17.2)

Hospital teaching status Nonteaching 5322 (42.6)Teaching 7173 (57.4)

Hospital size (bed count) Small 1476 (11.8)Medium 3397 (27.2)Large 7623 (61)

OOOO ORIGINAL ARTICLE

Volume 117, Number 4 Allareddy and Elangovan 473

5.8% of all hospitalizations were uninsured. Informa-tion on race was available for 9704 hospitalizations.The categories included white (51%), black (13.6%),Hispanic (26%), Asian/Pacific Islander (3.8%), NativeAmerican (0.3%), and other/mixed races (5.3%). Wefound that 85.4% were admitted to hospitals on anemergency or urgent basis.

A total of 29 different comorbid conditions wereexamined from the secondary diagnoses fields using theNIS severity index measures. The distribution of co-morbid conditions is summarized in Table II. Of allhospitalizations, 78.5% had at least 1 comorbid condi-tion. The 5 most frequent comorbid conditions werefluid and electrolyte disorders (65%), chronic pulmo-nary disease (9%), hypertension (8.4%), deficiencyanemias (6%), and weight loss (4.1%).

Disposition status after hospitalization is summa-rized in Table III. Of all hospitalized patients, 93.4%were discharged routinely, 0.7% were transferredto another short-term hospital, 2.3% were transferredto long-term care facilities, 2.8% were transferred tohome health care, and 0.3% were discharged againstmedical advice. A total of 68 patients died inhospitals.

Hospital characteristics are presented in Table IV.Hospitals located in the southern regions comprised41.3% of all hospitalizations. Teaching hospitals(57.4%) and large hospitals (high bed count) (61%)treated the majority of hospitalizations.

Length of stay and hospitalization charges (infla-tion-adjusted to 2010 dollars) over the study period aresummarized in Table V. The mean length of stayduring the study period was 3.9 days, and the totalnumber of hospitalization days across the entireUnited States was 49 472 days. The mean length ofstay was consistent annually across the time period,ranging from 3.7 days in 2008 to 4.1 days in 2005. Themean hospitalization charge over the entire studyperiod was $18 551. The total hospitalization chargesacross the entire United States from 2004 to 2010 were$229.4 million. The mean hospitalization chargesvaried from a low of $16 171 in 2008 to a high of$19 936 in 2006.

Table V. Hospitalization charges and length of stay after hospitalization due to herpetic gingivostomatitis

Outcome Measure 2004 2005 2006 2007 2008 2009 2010

Hospitalizationcharges in USdollars (inflation-adjusted to 2010dollars)

Mean 17 888 18 270 19 936 20 524 16 171 19 265 18 398Standard error 1668 1832 2405 2421 1507 2244 1485Total

hospitalizationcharges acrossentire UnitedStates

33 956 622 33 367 636 32 987 218 32 871 036 31 010 224 31 522 678 33 710 923

Length of stay inhospital (d)

Mean 3.9 4.1 4 3.9 3.7 3.9 3.9Standard error 0.18 0.27 0.21 0.23 0.16 0.23 0.14Total

hospitalizationdays acrossentire UnitedStates

7622 7631 6742 6439 7306 6463 7268

Total hospitalization days across the entire study period is 49 472.Total hospitalization charges across the entire study period is $229.4 million.

ORAL MEDICINE OOOO

474 Allareddy and Elangovan April 2014

DISCUSSIONTo our knowledge, this is the first study to explore theprevalence trends of hospitalizations due to HGS overa defined time period. Our analyses indicate that closeto 12 536 hospitalizations were projected to be due toHGS during the study period (2004 to 2010). As ex-pected, the study identified that most of the hospital-izations occurred in children (mean age, 15.6 years),with slightly more female patients than male patients.Close to 80% of the patients exhibited at least 1 co-morbid condition at the time of the visit, with fluidand electrolyte imbalance being the most commonamong patients hospitalized. The total hospitalizationcharges due to HGS across the entire United Statesfrom 2004 to 2010 were $229.4 million. In terms ofthe trend, both the number of hospitalizations per yearand their associated cost per year remained constantthroughout the study period. Some of these results areconsistent with our previous investigation that lookedat the prevalence of ED visits due to HGS.8 We re-ported that the average age of patients visiting an EDfor HGS in 2007 was 16, with slightly more femalepatients than male patients.8 The majority of patientswho visited an ED presented with comorbid condi-tions, with fever and electrolyte imbalances being thecommon ones.8

There are only a handful of studies in dentistry thathave used NIS, the largest all-payer in-hospital ad-missions data set, to examine the estimates of hospi-talizations that occurred due to oral conditions.10,11 Ina recent publication, Lee et al.10 looked at the esti-mates of hospitalizations attributed to common oralconditions in the United States, including the 2 mostcommons ones, dental caries and periodontal disease.The study found the total charge for hospitalizationsdue to oral conditions in the United States to be closeto $1.2 billion in 2008, confirming the substantial

utilization of resources to treat oral conditions inhospital settings.11

There are several factors that influence the outcomesamong hospitalized patients, such as complicationsarising from the condition of interest, comorbid con-ditions, iatrogenic complications, medical errors, and ahost of structural attributes of hospitals.12 Of these,comorbid conditions likely played a significant rolewith regard to HGS. From the results of the analysis, itis clear that several of the patients with HGS presentedwith at least 1 comorbid condition. Owing to severepain associated with this condition and the associatedintolerability to liquids, fluid and electrolyte imbalanceis extremely common in patients with HGS, and ourstudy further confirms this clinical observation.3

Considering that most of these patients were veryyoung and had fluid and electrolyte abnormalities,dehydration likely played a role in the caregiver’s de-cision to hospitalize these patients. Unfortunately, theinformation regarding the type of treatment rendered tothese patients during their hospital stay was not avail-able in NIS for analysis.

In the present study, close to 85.4% were admitted tohospitals on an emergency or urgent basis. Hospitali-zations after ED visits are very common for several oralconditions, which leads to significant utilization of re-sources.13,14 It is clear from recent studies that patientswith limited or no insurance coverage visit emergencydepartments for oral conditions such as dental caries orperiodontal disease more often than patients who haveadequate insurance coverage.13,14 It is also known thatpatients with private dental insurance visit the dentistmore often for routine diagnostic or preventive dentalcare than patients with public coverage.15 In this study,we noticed that during the study period, close to 45% ofthe hospitalizations occurred in patients who were un-der Medicaid insurance programs. These results, along

OOOO ORIGINAL ARTICLE

Volume 117, Number 4 Allareddy and Elangovan 475

with previous publications on this topic, clearly point tothe worrying trend that more and more patients withgovernment-sponsored insurance programs are notgetting routine dental care and end up going to hospitalsfor their oral health needs.13,14 Interestingly, aftergovernment-aided insurance, a significant number ofpatients who presented for hospitalization had privateinsurance coverage.

With regard to hospital locality, we noticed that aplurality of hospitalizations occurred in the southernregions of the United States. Socioeconomic levels andthe type of insurance coverage offered differgeographically, which could affect the geographicvariations in ED visits and hospitalizations.16 Thenumber and size of hospitals in the region and the factthat NIS uses sample data from only select states couldhave led to either overestimation or underestimation ofthe prevalence in a specific locality.

This assessment has the following limitations. Thisretrospective assessment relies heavily on ICD-CMcodes and on diagnoses made by health care providerswho may not be well versed in oral conditions. The NISdatabase does not provide information on virologicassays or other laboratory values. The diagnosis wasobtained from the primary diagnosis field in the data-base, which refers to the primary reason for hospitali-zation. We are unsure if HGS diagnosis was made byclinical diagnosis or virologic assay. Though the base-line demographic characteristics are captured accu-rately, the diagnosis, the severity of the condition, andthe procedures carried out in the hospitals to diagnoseand treat HGS may not be captured accurately.Although HGS is the primary diagnosis in these pa-tients, secondary comorbid conditions such as dehy-dration also necessitate substantial hospital resourceusage. Charges associated with hospitalizations, such astravel to and from the hospital, or the charges associatedwith medications after the stay are not included, andtherefore the total charge will be much more than whatwe report here. Additionally, although we used thelargest publicly available all-payer hospitalizationdatabase, it is important to remember that the finalnumbers are projected estimates based on data from astratified sample of hospitals throughout the UnitedStates.

This study, along with several of our earlier in-vestigations, further reinforces the importance oftraining medical professionals to diagnose, treat, orrefer patients with common oral conditions such asHGS. More and more dental institutions are adopting aninterprofessional educational curriculum, which willimprove communication among different branches ofmedicine and dentistry to deliver the best possible careto patients. On a long-term basis, improving access todental care for patients is crucial to overcome this

problem. It is equally important to promote preventiveprograms, such as a “dental home” for children, whichwill allow for early intervention and prevention ofcommon childhood oral infections, including thosecaused by HSV.17

CONCLUSIONSThe present study used the largest publicly availableall-payer hospital discharge data set in the United Statesto provide estimates of hospitalizations primarilyattributed to HGS. During the 7-year study period, atotal of 12 536 hospitalizations were attributable toHGS. Close to 80% of admitted patients had a comor-bid condition. A substantial amount of hospital re-sources was used to treat these patients in inpatientsettings.

REFERENCES1. Whitley RJ, Roizman B. Herpes simplex virus infections. Lancet.

2001;357:1513-1518.2. Siegel D, Golden E, Washington AE, et al. Prevalence and cor-

relates of herpes simplex infections: the population-based AIDSin Multiethnic Neighborhoods Study. JAMA. 1992;268:1702-1708.

3. Miller CS, Redding SW. Diagnosis and management of orofacialherpes simplex virus infections. Dent Clin North Am. 1992;36:879-895.

4. Scully C. Orofacial herpes simplex virus infections: current con-cepts in the epidemiology, pathogenesis and treatment, and dis-orders in which the virus may be implicated. Oral Surg Oral MedOral Pathol. 1989;68:701-710.

5. Arduino PG, Porter SR. Herpes simplex virus type 1 infection:overview on relevant clinico-pathologic features. J Oral PatholMed. 2008;37:107-121.

6. Siegel MA. Diagnosis and management of recurrent herpes sim-plex infections. J Am Dent Assoc. 2002;133:1245-1249.

7. Bradley H, Markowitz LE, Gibson T, McQuillan GM. Seropre-valence of herpes simplex virus types 1 and 2dUS, 1999-2010[published online October 16, 2013]. J Infect Dis.

8. Elangovan S, Karimbux NY, Srinivasan S, Venugopalan SR,Eswaran SV, Allareddy V. Hospital-based emergency departmentvisits with herpetic gingivostomatitis in the US. Oral Surg OralMed Oral Pathol Oral Radiol. 2012;113:505-511.

9. HCUP Nationwide Inpatient Sample (NIS). Healthcare Cost andUtilization Project (HCUP): 2004-2010. Rockville, MD: Agencyfor Healthcare Research and Quality. Available at: http://www.hcup-us.ahrq.gov/nisoverview.jsp. Accessed October 21, 2013.

10. Lee MK, Dodson TB, Nalliah RP, Karimbux NY, Allareddy V.Nine-year trend analysis of hospitalizations attributed to oral andoropharyngeal cancers in the US [published online April 6, 2013].Oral Surg Oral Med Oral Pathol Oral Radiol.

11. Allareddy V, Kim MK, Kim S, et al. Hospitalizations primarilyattributed to dental conditions in the US in 2008. Oral Surg OralMed Oral Pathol Oral Radiol. 2012;114:333-337.

12. Allareddy V, Konety BR. Characteristics of patients and pre-dictors of in-hospital mortality after hospitalization for head andneck cancers. Cancer. 2006;106:2382-2388.

13. Elangovan S, Nalliah R, Allareddy V, Karimbux NY,Allareddy V. Outcomes in patients visiting hospital EDs dueto periodontal conditions in US. J Periodontol. 2011;82:809-819.

ORAL MEDICINE OOOO

476 Allareddy and Elangovan April 2014

14. Nalliah RP, Allareddy V, Elangovan S, Karimbux N, Allareddy V.Hospital based emergency department visits attributed to dentalcaries in the US in 2006. J Evid Based Dent Pract. 2010;10:212-222.

15. AHRQ Web site. Available at: http://meps.ahrq.gov/data_files/publications/cb17/cb17.pdf. Accessed November 25, 2013.

16. Liu T, Sayre MR, Carleton SC. Emergency medical care: types,trends, and factors related to nonurgent visits. Acad Emerg Med.1999;6:1147-1152.

17. Nowak AJ, Casamassimo PS. The dental home: a primary careoral health concept. J Am Dent Assoc. 2002;133:93-98.

Reprint requests:

Veerasathpurush AllareddyDepartment of OrthodonticsCollege of DentistryUniversity of Iowa801 Newton RdIowa City, IA 52242USAVeerasathpurush-Allareddy@uiowa.edu; sathpurush@gmail.com