Percutaneous Nephrolithotomy Use Is Increasing in the United States: An Analysis of Trends and Complications

Abstract

Purpose:

To report the use and complication rates of percutaneous nephrolithotomy (PCNL) performed in the United States between 1998 and 2009.

Patients and Methods:

The Nationwide Inpatient Sample database was analyzed from 1998 to 2009 to identify all PCNL cases performed in adults ≥18 years old. Descriptive statistics were used for potential covariates: Demographics, comorbidities, academic/community hospital, rural/urban location, and U.S. geographic region. Common complications encoded by International Classification of Diseases-9 codes after PCNL were reported over time, and those found to be statistically significant were evaluated in the multivariate regression. Linear regression was used to analyze surgical trends. Multivariate regression was performed to identify covariates that predicted any surgical complication.

Results:

The use of PCNL among inpatients increased significantly from 15 to 27 surgeries/100,000 discharges between 1998 and 2009 (P<0.001), and this increase was seen in all geographic regions of the United States. The increase in adoption of PCNL was accompanied by an increase in complications (14% to 19%, P<0.001). Higher comorbidity (Charlson ≥3) was the strongest predictor of complications in multivariate analysis (odds ratio=2.22, P<0.001).

Conclusions:

This is the first study to demonstrate an increase in PCNL use in the United States over the last decade. While there was an increase in surgical complications during this same period, the complication rate found reported is commensurate with other international reports. PCNL is safe and use of percutaneous surgery in the United States will most likely continue to increase.

Introduction

T he prevalence of urinary stone disease is increasing in the United States with estimates that U.S. residents have a 10% to 15% chance of receiving a diagnosis of a kidney stone during their adult lives.¹ Fortunately, technologic advances over the past 30 years have offered nephrolithiasis patients several minimally invasive surgical options for those stones that cannot be managed conservatively. In particular, the introduction of percutaneous nephrolithotomy (PCNL) has transformed open kidney stone surgery into a rare phenomenon. PCNL now clearly has a role for large renal calculi >2 cm as described by European Association of Urology and the American Urological Association guidelines.^2,3 Several ostensible advances in PCNL explain this adoption: Improved equipment, surgeon familiarity with the procedure, refinements in surgical technique, and, most importantly, attainment of superior stone-free rates compared with shockwave lithotripsy (SWL) or ureteroscopy (URS).⁴

To determine if the use of PCNL was increasing over time nationally, we analyzed the Nationwide Inpatient Sample (NIS). We hypothesized that the likelihood of undergoing PCNL varied as a function of demographics, regionalization, and comorbidity. We also hypothesized that certain metrics, particularly complications, would decrease over time, contributing to the adoption of PCNL.

Patients and Methods

The NIS database is part of the Healthcare Cost and Utilization Project, sponsored by the Agency for Healthcare Research and Quality. The NIS is the largest all-payer inpatient care database in the United States, containing data on more than seven million hospital stays annually from approximately 1000 hospitals, constituting a 20% stratified sample of all U.S. hospitals. All data were weighted using discharge level values, based on the relative proportion of the total U.S. hospital patient population accounted for by that record, to produce 100% national estimates. The NIS includes charge information for all patients, regardless of payer, including patients covered by Medicare, Medicaid, private insurance, and the uninsured. Inpatient stay records include clinical, hospital, and resource use information typically available from discharge abstracts. Discharges from 1998 through 2009 were included in the analysis.

We identified adult patients (≥18 years old) with International Classification of Diseases (ICD)-9 procedural codes for PCNL (55.04 or 55.03/55.21) from 1998 through 2009. Patients undergoing a PCNL during their hospital stay, including those patients needing a second-look PCNL, were all captured using these ICD-9 codes and included in the analysis; patients who needed a subsequent SWL or URS during their inpatient hospital stay were excluded. Patient age, race, sex, treatment year, geographic region of treatment (Northeast, South, Midwest, and West), Charlson Comorbidity Index (CCI) score, comorbidities (hypertension [HTN], diabetes mellitus [DM], obesity), urban/rural location, teaching/community, and length of stay were abstracted.

A list of common complications was compiled. The composite of all complications—called “any complication”—was created. Multivariate analysis using binary logistic regression was performed to predict “any complication”; variables that were P<0.2 in univariate testing, or were determined to be clinically relevant were included in model development, with the final model including only those variables that retained significance (at P<0.05 level). Collinearity precluded use of certain variables (HTN, DM, obesity with CCI score). Chi-square, analysis of variance, and Student t tests (using Bonferroni correction for intergroup comparisons) were performed depending on whether the variable was continuous or categorical. Prevalence over time was examined using simple linear regression to produce β coefficients (average change over time). We determined which covariates increased over time using the chi-square test. All testing was two-tailed, with probability of type I error alpha=0.05, a priori. SPSS version 17.0 software (SPSS, Inc., Chicago, IL) was used for all statistical analyses.

Results

Approximately 83,676 hospital discharges (weighted value) were analyzed from 1998 through 2009. The use of PCNL increased significantly, from 15 to 27 surgeries/100,000 discharges, P<0.001 (Fig. 1). Although use of PCNL significantly increased in all regions of the country, the adoption increased most in the Midwest (β=136 surgeries/year) and least in the Northeast (β=45/year) (Fig. 2). More than 86% of patients undergoing PCNL had renal calculi, while 14% had ureteral calculi. Approximately 62.1% of all PCNL cases were performed at academic teaching hospitals (Table 1).

FIG. 1.

Inpatient percutaneous nephrolithotomies (per 100,000 discharges) over time.

FIG. 2.

Regional use of percutaneous nephrolithotomy (PNL) over time. squares: Northeast; triangles: Midwest; Xs: South; circles: West.

Table 1.

Patient and Hospital Characteristics

	PCNL (n=83,675)
Mean age±SE, year	53±0.2
Sex
Male	40,656 (48.8%)
Female	42,720 (51.2%)
Race
Caucasian	47,721 (57%)
Other	35,954 (43%)
Hypertension	31,621 (37.8%)
Diabetes	14,907 (17.8%)
Obesity	7563 (9%)
Charlson Comorbidity Index
0–2	80,217 (95.9%)
≥3	3458 (4.1%)
Teaching hospital	51,804 (62.1%)
Region
Northeast	16,258 (19.4%)
Midwest	20,710 (24.8%)
South	31,724 (37.9%)
West	14,982 (17.9%)
Location
Rural	4853 (5.8%)
Urban	78,568 (94.2%)
Other	8235 (9.8%)
Median length of stay (IQR), days	3 (2 to 4)

PCNL=percutaneous nephrolithotomy; SE=standard error; IQR=interquartile range.

The mean age of patients undergoing PCNL was 53 years, and 51.2% of patients were female (Table 1). HTN, DM, and obesity were observed in 38%, 18%, and 9% of the entire cohort, and the proportion of patients with DM, HTN, and obesity undergoing PCNL increased significantly over time, P<0.001. The obesity rates varied by sex with women nearly twice as likely to be obese than men (11.1% female vs 6.9% males, P<0.001). Approximately 4.1% of patients who underwent PCNL had a CCI score of ≥3. Similarly, the proportion of patients with CCI ≥3 undergoing PCNL also increased significantly over time, P=0.048. The median hospital length of stay was 3 days, and length of stay significantly decreased over time from 3 to 2 days, P<0.001.

In 13,020 (15.6%) patients, “any complication” was seen, and the three most common reported complications were urosepsis/urinary tract infection (8.4%), blood transfusions (4.4%), and sepsis (2.4%) (Table 2). The overall complication rate significantly increased from 14.1% to 18.5% between 1998 and 2009 (Fig. 3); specifically, rates of pneumothorax (P=0.031), transfusion (P<0.001) and sepsis (P<0.001) increased over time. Because complications increased rather than decreased as we originally hypothesized, we performed a subanalysis to better understand which demographic and disease factors may account for rising complications. Multivariate analysis demonstrated that patients with CCI ≥3 were 2.22 times more likely to have “any complication” (Table 3). Females were at increased risk of having “any complication.” PCNLs performed in the South or West and urban centers were associated with an increased chance of complication. The complication rates in rural (720/4853, 14.8%) and urban (12,227/78,568, 15.6%) hospitals as well as teaching (7,895/51,803, 15.2%) and nonteaching (5,052/31,617, 16%) hospitals were compared and no differences were observed, P=0.54 and P=0.33, respectively.

FIG. 3.

Complication rates over time.

Table 2.

Complication Rates and Changes Over Time

Complications (1998–2009)		Change over time P value
Any complication	13,020 (15.6%)	<0.001
Pneumothorax	257 (0.3%)	0.031
Transfusion	3,713 (4.4%)	<0.001
Flank cellulitis	91 (0.1%)	0.903
Sepsis	2,037 (2.4%)	<0.001
Pulmonary embolism	182 (0.2%)	0.94
Deep vein thrombosis	220 (0.3%)	0.844
Pneumonia	66 (0.1%)	0.799
Renal vascular injury	89 (0.1%)	0.883
Vascular injury	806 (1%)	0.72
Urosepsis/urinary tract infection	7,038 (8.4%)	0.628
Pyelonephritis	1,515 (1.8%)	0.241
Death	219 (0.3%)	0.564

Table 3.

Multivariate Analysis for Factors Associated with “Any Complication” ^*

		95% CI
Variables	OR	Lower	Upper	P value
Age, cont	1.022	1.021	1.024	<0.001
Female	2.144	2.065	2.226	<0.001
Charlson (3+, 0–2=ref)	2.224	2.091	2.366	<0.001
Hospital region (NE=ref)
Midwest	1.048	0.957	1.147	0.314
South	1.298	1.193	1.413	<0.001
West	1.317	1.191	1.457	<0.001
Urban (rural=ref)	1.179	1.088	1.278	<0.001
Year, cont	1.072	1.060	1.084	<0.001

Outcome=any complication (1998–2009).

OR=odds ratio; CI=confidence interval.

Discussion

Recent work by the Clinical Research Office of the Endourological Society (CROES) has confirmed that PCNL use is now widespread through the industrialized world. This project by CROES revealed that more than 5800 percutaneous procedures have been performed between November 2007 and December 2009.⁵ In fact, these numbers are arguably very conservative, because the database only abstracts a few academic centers in the United States and academic and community centers associated with the Endourological Society. Including data from all academic and community hospitals in these countries would produce a much higher prevalence of percutaneous surgery. By using the NIS, we were able to estimate national numbers for all inpatient percutaneous procedures performed at both academic and community hospitals. Our analysis confirmed that the global trend of increased use of PCNL also occurred in the United States between 1998 and 2009.

Moreover, our data illustrate the use of PCNL has almost doubled during this period. This analysis contradicts previous reports using Medicare datasets from a slightly earlier period (1992–1998), which demonstrated stable use of PCNL.⁶ Medicare datasets represent a selection bias because they are unfortunately limited to older patients and those selected persons eligible for Medicare coverage. The increased adoption of PCNL is also demonstrated across all regions of the United States, confirming a uniform, nationwide trend. Why the adoption is more pronounced in the Midwest and South compared with the West and Northeast is unclear. Increased prevalence of nephrolithiasis in these areas could potentially explain the regional trends. Larger stone burden as well as surgeon and patient preferences are potential explanations, but this evidence is beyond the scope of NIS. Financial incentives also may play a role as has been surmised for other stone procedures in a recent report on outpatient surgery in Florida⁷ and regionalization of complex procedures to tertiary centers may also contribute to preferential choice of certain procedures.⁸

We had hypothesized that the adoption of this new technology might be because of improved outcomes such as declining complications. The data instead showed that complications significantly increased from 14.1% to 18.5% during this period. Although our prevalence of complications is consistent with a recent systematic review reporting a 23.3% complication rate for PCNL,⁹ we surmise several plausible explanations for this increase in complications.

We know that according to several studies, complications in low-volume centers are greater than high-volume centers.^10
–12 Therefore, the diffusion of PCNL during our 10-year period was possibly accompanied by a large proportion of complications at low-volume centers attempting to adopt the technology. The CROES data also illustrated that severity of complications was initially higher until 100 to 150 cases were achieved.¹⁰ In fact, severity of complication had a bimodal distribution with severity increasing after 150 cases. The data also demonstrated that probability of complication was proportional to stone burden. Taken together, it is conceivable that as centers became more comfortable with PCNL, stone burden and case volume increased—having an overall increase in complication effect. Indeed it can be inferred that stone burden most likely increased given the significant rise in complications typical of larger stones, such as transfusions, sepsis, and pneumothorax.¹³

Comorbidity is an intuitive covariate of PCNL complications. Tefekli and associates¹⁴ demonstrated that complications were 2.7 times more common in patients with DM and HTN. Patients with metabolic syndrome also needed more ancillary treatments after PCNL. Similarly, our study demonstrated that comorbidity (CCI ≥3) increased the risk of complications by 2.2. Increased prevalence of DM and HTN in patients undergoing percutaneous surgery over time in our cohort may thus account for the increased rate of complications reported from 1998 to 2009. The rising prevalence of metabolic syndrome (DM, HTN, obesity, hyperlipidemia) in the United States over the last decade¹⁵ and the association between metabolic syndrome and nephrolithiasis^16
–18 supports our supposition that these cocovariates may have contributed to the rise in PCNL complications.

Limitations of the NIS are inherent to this dataset: It relies on accurate physician coding, and it does not provide clinical data (eg, stone size) that would possibly provide insight into procedure choice. Another limitation of the NIS is that complications are not classified according to the Clavien-Dindo system,¹⁹ and thus direct comparisons of severity of complications with other large PCNL case series are not possible. There is a trend toward percutaneous surgery in an outpatient setting, and we admittedly are missing this data; however, this surgery is still predominantly an inpatient procedure. Notwithstanding the above limitations, this study is important in that it quantitatively confirms that percutaneous stone surgery has increased over the last decade in the United States.

Conclusions

In the largest U.S. NIS, PCNL use has increased by nearly twofold. Although there was a concomitant increase in complications, this perhaps is a natural phenomenon observed during diffusion of an established surgical technique to patients with more comorbidities at hospitals less experienced in performing this procedure. The proportion of complications is consistent with worldwide reports and supports the continued use of percutaneous surgery. Increased use will likely witness a plateau in complications, further justifying the advantages of percutaneous surgery.

Footnotes

Disclosure Statement

No competing financial interests exist.

Abbreviations Used

References

Scales

Jr , Smith

, Hanley

, Saigal

. Prevalence of kidney stones in the United States. Eur Urol, 2012; 62:160–165.

Preminger

, Assimos

, Lingeman

et al. Chapter 1: AUA guideline on management of staghorn calculi: Diagnosis and treatment recommendations. J Urol, 2005; 173:1991–2000.

Tiselius

, Ackermann

, Alken

et al. Guidelines on urolithiasis. Eur Urol, 2001; 40:362–371.

Kim

, Kuo

, Lingeman

. Percutaneous nephrolithotomy: An update. Curr Opin Urol, 2003; 13:235–241.

de la Rosette

, Assimos

, Desai

et al. The Clinical Research Office of the Endourological Society Percutaneous Nephrolithotomy Global Study: Indications, complications, and outcomes in 5803 patients. J Endourol, 2011; 25:11–17.

Pearle

, Calhoun

, Curhan

. Urologic diseases in America project: Urolithiasis. J Urol, 2005; 173:848–857.

Hollingsworth

, Krein

, Birkmeyer

et al. Opening ambulatory surgery centers and stone surgery rates in health care markets. J Urol, 2010; 184:967–971.

Morris

, Taub

, Wei

et al. Regionalization of percutaneous nephrolithotomy: Evidence for the increasing burden of care on tertiary centers. J Urol, 2006; 176:242–246.

Seitz

, Desai

, Häcker

et al. Incidence, prevention, and management of complications following percutaneous nephrolitholapaxy. Eur Urol, 2012; 61:146–158.

10.

Opondo

, Tefekli

, Esen

et al. Impact of case volumes on the outcomes of percutaneous nephrolithotomy. Eur Urol, 2012; 62:1181–1187.

11.

Duvdevani

, Razvi

, Sofer

et al. Third prize: Contemporary percutaneous nephrolithotripsy: 1585 procedures in 1338 consecutive patients. J Endourol, 2007; 21:824–829.

12.

de la Rosette

, Zuazu

, Tsakiris

et al. Prognostic factors and percutaneous nephrolithotomy morbidity: A multivariate analysis of a contemporary series using the Clavien classification. J Urol, 2008; 180:2489–2493.

13.

Srivastava

, Singh

, Suri

et al.

Vascular complications after percutaneous nephrolithotomy: Are there any predictive factors?

Urology, 2005; 66:38–40.

14.

Tefekli

, Kurtoglu

, Tepeler

et al.

Does the metabolic syndrome or its components affect the outcome of percutaneous nephrolithotomy?

J Endourol, 2008; 22:35–40.

15.

Mokdad

, Bowman

, Ford

et al. The continuing epidemics of obesity and diabetes in the United States. JAMA, 2001; 286:1195–1200.

16.

Lange

, Mufarrij

, Wood

et al. The association of cardiovascular disease and metabolic syndrome with nephrolithiasis. Curr Opin Urol, 2012; 22:154–159.

17.

Rendina

, Mossetti

, De Filippo

et al. Association between metabolic syndrome and nephrolithiasis in an inpatient population in southern Italy: Role of gender, hypertension and abdominal obesity. Nephrol Dial Transplant, 2009; 24:900–906.

18.

Jeong

, Kang

, Bang

et al. Association between metabolic syndrome and the presence of kidney stones in a screened population. Am J Kidney Dis, 2011; 58:383–388.

19.

Dindo

, Demartines

, Clavien

. Classification of surgical complications: A new proposal with evaluation in a cohort of 6336 patients and results of a survey. Ann Surg, 2004; 240:205–213.