Reasons for hospitalization and in-hospital mortality in adults with systemic sclerosis: Analysis of the National Inpatient Sample

Abstract

Objective:

Systemic sclerosis is an autoimmune condition with significant morbidity and mortality despite modern medical therapies. The goal of this investigation was to comprehensively analyze all reasons for hospitalization and in-hospital death of systemic sclerosis patients.

Methods:

We conducted a retrospective analysis of the adult systemic sclerosis hospitalizations from the 2016–2018 National Inpatient Sample. We included patients with a primary or secondary diagnosis of systemic sclerosis and compared them to the group without the disease. The incidence of inpatient death and total hospitalization charges were recorded along with the most frequent principal diagnoses for systemic sclerosis hospitalizations and mortality categorized into subgroups.

Results:

There were 94,515 adult systemic sclerosis hospitalizations recorded in the 2016–2018 National Inpatient Sample database. Systemic sclerosis patients had higher inpatient mortality compared to the non-systemic sclerosis group (4.5% vs 2.2%, respectively, p < 0.0001), were more likely to be female (84% vs 58%, p < 0.0001), had a longer mean length of stay (6.1 vs 4.7 days, p < 0.0001), and greater mean total hospital charges ($70,018 vs $53,556, p < 0.0001). Sepsis, unspecified organism (A41.9) was the most common principal diagnosis for both hospitalized and deceased systemic sclerosis patients. Cardiovascular diagnoses (21.9%) were the most common reasons for hospitalization and infectious (28%)—for in-hospital death.

Conclusion:

Our analysis of the National Inpatient Sample database from 2016 to 2018 showed that infections and cardiovascular diseases were a significant cause of morbidity and mortality among hospitalized systemic sclerosis patients. Sepsis was the most frequent specific diagnosis for both hospitalization and inpatient deaths. These results stress the importance of early recognition of life-threatening infections in this patient population.

Introduction

Systemic sclerosis (SSc) is a chronic, immune-mediated disease that damages the skin and internal organs by a combination of vascular disease, immune system activation, and fibrosis. ¹ It is traditionally classified into two groups, limited cutaneous systemic sclerosis (lcSSc) and diffuse cutaneous systemic sclerosis (dcSSc) based on the extent of the skin involvement and associated pattern of internal organ disease. Despite an overall improvement in survival and treatment methods, diagnosis of SSc carries significant morbidity and mortality, especially for the patients with dcSSc.^2,3 Although there are previous studies analyzing the most common reasons for hospitalizations and mortality for SSc patients, they either focused on specific disease subgroups, were based on an older database, or had a significantly smaller study sample.^{4 –10} The aim of this investigation was to use the 2016–2018 United States National Inpatient Sample (NIS) to comprehensively categorize all reasons for hospitalization and in-hospital death in patients admitted with a diagnosis of SSc. In addition, we aim to describe the economic burden of these hospitalizations and the most frequent admitting diagnoses.

Materials and methods

Data source

We conducted an analysis of adult SSc hospitalizations in 2016–018 using the NIS database (online at http://www.hcup-us.ahrq.gov). The NIS is the largest collection of inpatient data from all nonfederal acute care hospitals in the United States. The NIS is a 20% sample from each region of the country and represents more than 97% of the US population. Discharges were recorded and then weighted to ensure that they are nationally representative. In the NIS, diagnoses are divided into two separate categories: principal diagnosis and secondary diagnoses. The principal diagnosis is the main discharge International Classification of Diseases, 10th Revision, Clinical Modification (ICD-10) code for the hospitalization. Secondary diagnoses are the ICD-10 codes for any diagnoses other than the principal diagnosis. ¹¹ Institutional Review Board approval was not pursued because data in NIS are de-identified and publicly available but as per guidelines of the Healthcare Cost and Utilization Project (HCUP) no cells containing discharge records of a value of 1–10 were reported in either text or tables to protect patient privacy.

Inclusion criteria and study variables

We included hospitalizations in 2016, 2017, and 2018 NIS with a principal or secondary diagnosis of SSc. We used the ICD-10 code M34 to identify hospitalizations among patients with SSc (see Appendix 1 for list of codes). We excluded patients ⩽18 years old. The reference population consisted of all adult hospitalizations without SSc. We collected demographic characteristics, mean length of stay (LOS), total hospital charges in dollars, mortality, and ICD-10 diagnoses. The principal “reason for hospitalization” and “reason for in-hospital death” in patients with SSc were divided into 19 organ system/disease categories based on their principal ICD-10 hospital billing diagnosis. There are over 69,000 ICD-10 codes. Each code is composed of seven digits. The first three digits of ICD-10 codes indicate the disease category. The fourth to seventh digits add details. ICD-10 codes were grouped into organ system/disease categories by the first three digits as follows:

• Infectious diagnoses, codes A00–B99:

• Hematologic, Oncologic, and Neoplastic diagnoses, codes C00–D89;

• Endocrine diagnoses, codes, E00–E89;

• Psychiatric diagnoses, codes F01–F99;

• Neurologic diagnoses, codes G00–G99;

• Eye/Ear diagnoses, codes H00–H95;

• Cardiovascular diagnoses, codes I00–I99;

• Respiratory diagnoses, codes J00–J99;

• Gastrointestinal diagnoses, codes K00–K95;

• Skin diagnoses, codes L00–L99;

• Rheumatologic diagnoses, codes M00–M99; SSc M34 codes are a subset of the larger rheumatologic category;

• Genitourinary (GU) diagnoses, codes N00–N99;

• Obstetrics (OB) diagnoses, codes O00–O9A;

• Perinatal diagnoses, codes P00–P96;

• Congenital diagnoses, codes Q00–Q99;

• Other, codes R00–R99;

• Injury & poisoning, codes S00–T88;

• Accidents, codes V00–Y99;

• Health services, codes Z00–Z99.

Outcomes

The main outcomes studied were (a) the incidence of in-hospital death among SSc hospitalizations in the NIS in the period from 2016 to 2018 and (b) categorization for the “reasons for hospitalization” and “in-hospital death” by the principal discharge diagnosis. (c) identification of the most frequent principal ICD-10 diagnoses for SSc hospitalizations (d) quantification of the total charges of SSc hospitalizations, and (e) determine incidence of sepsis among SSc hospitalizations.

Statistical analysis

Analyses were performed using STATA, version 16 (StataCorp, Texas, USA). The patients with SSc were identified using ICD coding as described above. Descriptive statistics were used to compare patients in SSc and non-SSc groups. Logistic and linear regression were used for categorical and continuous outcomes, respectively. A p value of <0.05 was considered significant. The NIS was queried for conditions from each disease subgroup using ICD-10 codes and proportions among all hospitalizations were calculated. The most common ICD-10 codes for both principal discharge diagnosis and inpatient mortality were ranked using the STATA program.

Results

Patient characteristics and reasons for hospitalization

After weighting, there were 90,879,561 adult hospitalizations recorded in the combined 2016, 2017, and 2018 NIS database. Of those, 94,515 had either a principal or secondary ICD-10 code for SSc. Characteristics of SSc and non-SSc hospitalizations are displayed in Table 1. Relative to the reference population, hospitalized SSc patients were more likely to be female (84% vs 58%, p < 0.0001), and older (62.8 vs 57.9 years, p < 0.0001). The racial/ethnic categories of SSc hospitalizations were as follows: White (64.7%), African American (15.5%), Hispanic (11.2%), Asian or Pacific Islander (1.9%), Native American (0.8%), and other (2.5%). Compared to the reference population, SSc hospitalizations had a longer mean LOS (6.1 vs 4.7 days, p < 0.0001) and greater mean total hospital charges ($70,018 vs $53,556, p < 0.0001). Mortality was higher for SSc hospitalizations as well (4.5% vs 2.2%, p < 0.0001). The five most common principal ICD-10 diagnoses (Table 2) were sepsis unspecified organism A41.9, pneumonia J18.9, acute kidney failure unspecified N17.9, hypertensive heart disease with heart failure I11.0, and hypertensive heart and chronic kidney disease with heart failure and stage 1 through stage 4 chronic kidney disease, or unspecified chronic kidney disease I13.0.

Table 1.

Weighted descriptive characteristics of adult hospitalizations in NIS 2016–2018.

Hospitalization characteristics	Scleroderma hospitalizations (n = 94,515)	Non-scleroderma hospitalizations (n = 90,785,046)	p value
Women, number (%)	79,630 (84%)	52,390,112 (58%)	<0.0001
Age, mean in years	62.8	57.9	<0.0001
Age distribution, number (%)
Age 18–40 years	7060 (7.5%)	21,943,086 (24.2%)	<0.0001
Age 40–60 years	28,245 (30%)	21,673,207 (24%)	<0.0001
Age 60–80 years	48,105 (51%)	32,311,596 (35.6%)	<0.0001
Age > 80 years	11,105 (11.7%)	14,857,157 (16.4%)	<0.0001
Race number (%)
(a) White	61,130 (64.7%)	59,032,624 (65%)	0.419
(b) African American	14,675 (15.5%)	13,282,695 (14.6%)	0.004
(c) Hispanic, number	10,590 (11.2%)	9,772,919 (10.7%)	0.114
(d) Asian or Pacific Islander	1770 (1.9%)	2,407,328 (2.7%)	<0.0001
(e) Native American	750 (0.8%)	551,465 (0.6%)	0.002
(f) other	2360 (2.5%)	2,633,627 (2.9%)	0.005
Length of stay, mean days	6.1	4.7	<0.0001
Mortality, number (%)	4245 (4.5%)	2,016,089 (2.2%)	<0.0001
Total Charges, mean dollars	$70,018	$53,556	<0.0001

NIS: National Inpatient Sample.

Table 2.

Top principal ICD-10 diagnoses in scleroderma hospitalizations.

Top principal ICD-10 diagnoses in scleroderma hospitalizations	Weighted number of hospitalizations
Sepsis, unspecified organism A41.9	6865
Pneumonia J18.9	2495
Acute kidney failure unspecified N17.9	1675
Hypertensive heart disease with heart failure I11.0	1665
Hypertensive heart and chronic kidney disease with heart failure and stage 1 through stage 4 chronic kidney disease, or unspecified chronic kidney disease I13.0	1605
Acute and chronic respiratory failure with hypoxia J96.21	1460
Systemic sclerosis, unspecified M34.9	1405
Non-ST elevation (NSTEMI) myocardial infarction I21.4	1305
Other secondary pulmonary hypertension I27.2	1085
Systemic sclerosis with lung involvement M34.81	1055

ICD: International Classification of Diseases.

Reasons for hospitalization categorized by the principal diagnosis

The reasons for hospitalization among SSc patients by their principal discharge diagnosis category were as follows (Table 3 and Figure 1): cardiovascular (21.9%), gastrointestinal (14%), respiratory (13.1%), infectious (10.9%), rheumatologic (10.3%), injury and poisoning (8.1%), hematology/oncology (4%), genitourinary (3.8%), other (3%), endocrine (3%), skin (2.5%), neurologic (1.9%), psychiatry (1.3%), health services (0.9%), obstetrics (0.8%), eye and ear (0.2%), congenital (0.1%), perinatal (0%), and accidents (0%).

Table 3.

Reasons for hospitalization by principal diagnosis category NIS 2016–2018.

ICD-10 diagnosis category	Weighted number (%) of hospitalizations in scleroderma patients (n = 94,515)	Weighted number (%) of hospitalizations in non-scleroderma patients (n = 90,785,046)	P value
Cardiovascular	20,680 (21.9%)	14,744,851 (16.2%)	<0.0001
Gastrointestinal	13,190 (14%)	9,027,411 (10%)	<0.0001
Respiratory	12,410 (13.1%)	7,300,785 (8%)	<0.0001
Infections	10,260 (10.9%)	7,086,858 (7.8%)	<0.0001
Rheumatologic	9770 (10.3%)	6,706,611 (7.4%)	<0.0001
Injury/poison	7670 (8.1%)	7,815,403 (8.6%)	0.021
Hematology/oncology	3735 (4%)	4,851,318 (5.3%)	<0.0001
Genitourinary	3620 (3.8%)	4,289,778 (4.7%)	<0.0001
Other	2910 (3%)	2,705,179 (3%)	0.428
Endocrine	2875 (3%)	3,753,348 (4.1%)	<0.0001
Skin	2370 (2.5%)	1,787,114 (2%)	<0.0001
Neurologic	1810 (1.9%)	2,340,260 (2.6%)	<0.0001
Psychiatry	1260 (1.3%)	5,253,937 (5.8%)	<0.0001
Health services	925 (0.9%)	938,945 (1%)	0.507
Obstetrics	735 (0.8%)	11,869,280 (13.1%)	<0.0001
Eye/ear	180 (0.2%)	163,850 (0.2%)	0.758
Congenital	115 (0.1%)	115,375 (0.1%)	0.834
Perinatal	0 (0%)	765 (0.0008%)	–
Accidents	0 (0)	535 (0.0006%)	–

NIS: National Inpatient Sample; ICD: International Classification of Diseases.

Figure 1.

Top 10 causes of hospitalizations of systemic sclerosis patients.

Mortality

A total of 4245 (4.5%) hospitalized patients with a principal or secondary ICD-10 code for SSc suffered in-hospital death. The five most common specific principal ICD-10 diagnoses (see Table 4) in SSc hospitalizations resulting in in-hospital death were sepsis, unspecified organism A41.9, acute and chronic respiratory failure with hypoxia J96.21, acute respiratory failure with hypoxia J96.01, SSc with lung involvement M34.81, and hypertensive heart and chronic kidney disease with heart failure and stage 1 through stage 4 chronic kidney disease, or unspecified chronic kidney disease I13.0.

Table 4.

Top principal ICD-10 diagnoses in scleroderma patients with in-hospital mortality.

Top principal ICD-10 diagnoses in scleroderma with in-hospital mortality	Weighted number of hospitalizations
Sepsis, unspecified organism A41.9	950
Acute and chronic respiratory failure with hypoxia J96.21	205
Acute respiratory failure with hypoxia J96.01	135
Systemic sclerosis with lung involvement M34.81	120
Hypertensive heart and chronic kidney disease with heart failure and stage 1–4 chronic kidney disease, or unspecified chronic kidney disease I13.0	105
Other secondary pulmonary hypertension I27.2	100
Non-ST elevation (NSTEMI) myocardial infarction I21.4	90
Hypertensive heart disease with heart failure I11.0	80
Pneumonia J18.9	75
Systemic sclerosis, unspecified M34.9	70

ICD: International Classification of Diseases.

Reasons for in-hospital death categorized by the principal diagnosis

The reasons for in-hospital death by principal discharge diagnosis category were as follows (Table 5): infectious (28%), cardiovascular (23%), respiratory (17%), rheumatologic (7.9%), gastrointestinal (7.5%), injury and poisoning (4.2%), hematology/oncology (4.2%), genitourinary (2.6%), endocrine (1.5%), other (1.3%), health services (0.9%), neurologic (0.8%), skin (0.6%), obstetrics (< 1%), eye and ear (0%), psychiatry (0%), perinatal (0%), congenital (0%), and accidents (0%).

Table 5.

Reason for in-hospital mortality by principal diagnosis category NIS 2016–2018.

ICD-10 principal diagnosis category	Weighted number (%) of in-hospital deaths in scleroderma patients (n = 4245)	Weighted number (%) of in-hospital deaths in non-scleroderma patients (n = 2,016,089)	P value
Infectious	1185 (28%)	593,050 (29.4%)	0.344
Cardiovascular	990 (23%)	477,080 (23.7%)	0.808
Respiratory	720 (17%)	267,435 (13.3%)	0.002
Rheumatologic	355 (7.9%)	12,990 (0.6%)	<0.0001
Gastrointestinal	320 (7.5%)	131,615 (6.5%)	0.227
Injury/poison	180 (4.2%)	172,300 (8.5%)	<0.0001
Hematology/oncology	180 (4.2%)	161,355 (8%)	<0.0001
Genitourinary	110 (2.6%)	51,865 (2.6%)	0.972
Endocrine	65 (1.5%)	30,450 (1.5%)	0.960
Other	55 (1.3%)	25,010 (1.2%)	0.885
Health services	40 (0.9%)	42,310 (2.1%)	0.021
Neurologic	35 (0.8%)	35,600 (1.8%)	0.042
Skin	25 (0.6%)	6310 (0.3%)	0.155
Obstetrics	⩽10 (<1%)	1175 (0.05%)	0.481
Eye/ear	0 (0%)	165 (0.008%)	–
Psychiatry	0 (0%)	4860 (0.2%)	–
Perinatal	0 (0%)	35 (0.002%)	–
Congenital	0 (0%)	785 (0.03%)	–
Accidents	0 (0%)	0 (0%)	–

ICD: International Classification of Diseases.

Sepsis and SSc

Sepsis, unspecified organism A41.9 was the most common principal diagnosis for both hospitalized and deceased SSc patients. Our analysis also showed that 8995 (9%) of the hospitalizations and 1155 (27%) deaths had some type of sepsis code (see Appendix 1) listed as the principal diagnosis.

Discussion

SSc is known to have a worse prognosis than most autoimmune conditions.^2,3 Even though management of life-threatening manifestations such as scleroderma renal crisis or pulmonary arterial hypertension (PAH) has greatly improved over the past years, morbidity and mortality are still high among the population affected by the disease. The complexity of patients` condition and burden of other complications severely affecting the quality of life cause substantial difficulties in the management of SSc patients, ² as well as remarkable financial strain. To clarify the morbidity and mortality of SSc patients on a national level, we completed a comprehensive review of hospitalizations using the three most recent releases of the NIS database. Prior NIS studies either reported on specific diagnoses or had smaller cohorts.^{4 –7} Other observational studies were mostly based on an outpatient rather than inpatient patient population and also involved a smaller sample.^12,13 Our methodology provides a more complete picture of SSc hospitalizations and inpatient mortality by categorizing hospitalizations and deaths by the initial three digits of the principal ICD-10 diagnosis code.

Overall, patients with SSc have significantly higher mortality comparing to the general population^3,12,14 with a reported standardized mortality ratio (SMR) of 2.72 in the recent meta-analysis from 2014, ³ and SMR of 1.5–7.2 in the international meta-analysis from 2005. ¹⁴ Our study also supports these findings showing significantly higher mortality in the SSc population comparing to the non-SSc group among hospitalized patients (4.5% vs 2.2%, respectively, with p value < 0.0001). In-hospital mortality rates for SSc hospitalizations in several prior studies were comparable to our results,^{5 –7,9} but others reported rates as high as 16%–17%.^8,10 NIS analysis conducted by Singh et al. also proved that SSc patients admitted with serious infections had a higher risk of inpatient death compared to patients without SSc (9% vs 6.2%, respectively), as well as longer LOS and higher hospital charges. ⁴ To provide additional information, in the previous investigations renal, pulmonary, and cardiac involvement were found to be important predictors of mortality,^3,5,6,14 as well as the presence of anti-topoisomerase I antibodies, ¹⁴ and aspiration.^6,9 It is worth mentioning that the financial burden of SSc hospitalizations is significant as well and has been shown to be up-trending in our investigation compared to prior results. ⁴

Our study showed that infections remain a crucial aspect of morbidity and mortality of hospitalized SSc patients, accounting for 28% of all in-hospital deaths and 10.9% of the reasons for hospitalizations overall. This information appears to be in concordance with some of the prior investigations. Singh and Cleveland ⁴ conducted an NIS analysis of hospitalizations of SSc patients from 1998 to 2016. It is worth mentioning that the study was related only to admissions for serious infections and other reasons for hospitalizations were not investigated. The infectious causes were divided into five categories: opportunistic, skin and soft tissue, urinary tract infections, pneumonia, and sepsis. The study showed that during that period, the most common serious infection among hospitalized SSc patients was pneumonia; however, it was surpassed by sepsis in 2013–2014 which by 2015–2016 became 1.8 times more frequent than pneumonia. ⁴ Similarly, in our analysis sepsis was the most common principal diagnosis for both hospitalized and deceased SSc patients. This trend can be partially attributed to increased sensitivity of sepsis coding rather than truly increased number of sepsis hospitalizations; ¹⁵ nevertheless, it has to be taken into serious consideration. A more recent analysis of NIS 2012–2013 by Poudel et al. yielded comparable results. ⁶ The most common reason for both inpatient death and hospitalizations, in general, were infections (32.7% and 17.4%, respectively). Their study limitations included fewer disease categories, a smaller sample, and an older database. ⁶ An analysis of SSc Thai population also demonstrated that infections were the most common cause of hospitalizations and inpatient death, particularly due to bacterial pneumonia. ⁸ These results highlight the importance of keeping a high level of clinical suspicion for infectious complications in order to reduce morbidity and mortality among hospitalized SSc patients. This is especially important, given the fact that SSc patients seem to have worse hospital outcomes compared to the non-SSc population. It is also crucial to ensure SSc patients receive appropriate vaccinations.

With regards to cardiovascular disease, in our study, it was the most common reason for hospitalization of SSc patients by the principal diagnosis (21.9%) and also an important cause of inpatient mortality (23%, which was the second most common disease subgroup after infections). SSc patients are known to be more prone to cardiovascular complications.^1,2 According to prior studies, the main cause of death of SSc patients were cardiopulmonary complications.^3,12,13 Comparably, a previous NIS analysis performed by Chung et al. ⁵ investigated the HCUP data from 2002 and 2003 and found that cardiovascular diseases were the most common reason for both hospitalization (21.9%) and inpatient mortality (25.8%). With regards to the study limitations, their data only included information from 35 states in 2002 and 37 states in 2003 and contained only 14 secondary diagnoses. ⁵ The cardiovascular category encompasses PAH which is a well-known complication of SSc. Similar to the scleroderma renal crisis, there has been improvement in the outcomes of PAH related to implementation of modern therapies. A recent meta-analysis from 2021 showed a 36% reduction in the risk of clinical morbidity and mortality in connective tissue disease-associated PAH (CTD-PAH) treated with prostacyclin receptor agonists or endothelin receptor antagonists as compared to the control group; the survival rate of CTD-PAH patients also showed an increasing trend over time. ¹⁶ Another meta-analysis from 2021 investigated vascular disease in SSc patients, including digital ulcers, PAH, and scleroderma renal crisis and although there was no overall change in pooled prevalence of these three conditions over time, an improvement in 1-year PAH mortality and overall renal crisis mortality was observed. ¹⁷ It is important to remember that despite an overall improvement in the outcomes, PAH remains an important clinical problem in SSc patients. In our analysis, diagnosis of PAH was present in the top-10 most common reasons both for hospitalizations and inpatient mortality. A recent analysis of the Nationwide Readmission Database 2010–2015 demonstrated that even though the readmission rates during that time decreased for non-SSc-related PAH (23% to 20%, p < 0.001), it remained the same for SSc-related PAH (23% to 23%, p = 0.77). PAH patients with a diagnosis of SSc had higher mortality during both index hospitalization and readmissions compared to the non-SSc group. ¹⁸ To summarize, multidisciplinary patient care, collaboration between specialists, involvement of a cardiologist, and enhanced patient education are all likely important aspects needed to improve the outcomes of cardiovascular disease among SSc patients.

Not surprisingly, gastrointestinal (GI) disease is also an important cause of morbidity in the SSc population. Up to 90% of SSc patients experience some degree of GI symptoms, including gastroesophageal reflux, dysphagia, and diarrhea, among others. ¹ Some of the complications such as GI bleeding might be life-threatening. ¹ According to the results of our study, GI disease constituted 14% of the reasons for hospitalizations (second most common disease subgroup after cardiovascular); however, GI diagnoses were a less frequent cause of in-hospital death (7.5%). That might suggest that while GI symptoms tend to be more bothersome and prompt the patients to present to the hospital, they are less likely to result in death. Poudel et al. observed a similar trend with GI diseases accounting for 13.3% of all reasons for hospitalizations but only 4.8% of inpatient deaths. In that study, the opposite was noted for respiratory diseases which accounted for 10.9% of all hospitalizations but constituted 20% of deaths. ⁶ According to the investigation conducted by Chung et al., pulmonary fibrosis was the second most common principal diagnosis, and it also increased the odds of in-hospital death and prolonged the LOS. Respiratory failure was the second most common diagnosis in patients who died. ⁵ In our study, respiratory diagnoses were the third most common pathology for both hospitalization and inpatient death. Acute and chronic respiratory failure with hypoxia, acute respiratory failure with hypoxia, and SSc with lung involvement were the second, third, and fourth most common individual diagnoses for in-hospital deaths, respectively.

Historically, renal involvement has been one of the most common life-threatening complications of SSc. A recent decrease in deaths due to kidney involvement has been observed previously. ¹³ In our analysis, genitourinary diagnoses were only the eighth most common reason for both hospitalizations and inpatient mortality (3.8% and 2.6%, respectively). Overall improvement of survival in SSc patients over the past few decades has been primarily related to better treatment options for scleroderma renal crisis. ¹³

Our study’s major strength is the large sample size, characteristic of US hospitalizations. It is an up-to-date study comprehensively categorizing all reasons for hospitalization and in-hospital death that provides a novel insight into the current trend of SSc hospitalizations. It has several limitations. First, the study relies exclusively on coding, so it is not possible to verify clinical details. Nevertheless, the sensitivity and specificity of ICD coding for SSc have been shown to be high at 80.5% and 95%, respectively. ¹⁹ Second, this report reflects data on total SSc hospitalizations rather than on individuals; patients who were hospitalized multiple times could not be distinguished. The data regarding readmissions among SSc patients is scarce in the literature, but it is likely that patients with SSc have higher readmission rates than general population. A separate analysis using the Nationwide Readmissions Database could provide that valuable information in a future study. Third, categorization by the principal ICD billing codes can underrepresent some categories. For example, based on the initial 3-digit categorization, infectious arthritis is placed in the rheumatologic category and bacterial pneumonia is placed in the pulmonary category. Fourth, is worth mentioning that despite the sizable population included in the study, the results might not be representative of other countries and healthcare systems, especially when it comes to economic data. Last but not least, the NIS does not contain data regarding medications such as immunosuppressants, laboratory results and other disease-specific characteristics (such as duration or clinical course of the disease); thus it is not possible to include these variables in the analysis.

In summary, our analysis of the NIS database from 2016 to 2018 showed that infections and cardiovascular disease were a significant cause of morbidity and mortality among hospitalized SSc patients. The results of our study share some similarities with previous investigations related to SSc. Sepsis was the most frequent specific diagnosis for both hospitalization and inpatient deaths. These results stress the importance of early recognition of life-threatening infections and cardiopulmonary complications in this patient population.

Appendix 1

M34 Systemic sclerosis [scleroderma]

M34.0 Progressive systemic sclerosis

M34.1 CR(E)ST syndrome

M34.2 Systemic sclerosis induced by drug and chemical

M34.8 Other forms of systemic sclerosis

M34.81 Systemic sclerosis with lung involvement

M34.82 Systemic sclerosis with myopathy

M34.83 Systemic sclerosis with polyneuropathy

M34.89 Other systemic sclerosis

M34.9 Systemic sclerosis, unspecified

Sepsis codes

A40.x. . . . . . Streptococcal sepsis

A41.x. . . . . .Other sepsis (Staphylococcus, Hemophilus influenzae, anaerobes, Escherichia coli, other gram-negative organisms),

A02.01. . . . . . Salmonella sepsis

A32.7. . . . . . Listeria sepsis

A54.86. . . . . . Gonococcal sepsis

B37.7. . . . . . Candida sepsis

R65.20. . . . . . Severe sepsis without septic shock

R65.21. . . . . . Septic shock.