Professional Documents
Culture Documents
Measure
Measure
doi: 10.1093/ndt/gfq849
Advance Access publication 9 February 2011
CKD—thereby neglecting the chronicity which is inherent Table 1. The stages of CKD as outlined by the National Kidney
in this disease [9, 10]. Foundation Kidney Disease Outcomes Quality Initiative [4]
There has also been much debate regarding the best
Stage Clinical features GFR (mL/min/1.73m2)
method of measuring kidney function. The gold standard
of assessing kidney function is direct measurement of GFR I a
Kidney damage with normal 90
through 24-h urine collection. However, this is not practical or increased GFR
a
in many settings; consequently, equations are used to cal- II Kidney damage with a mild 60–89
culate eGFR in lieu of direct measurement. The two equa- decrease in GFR
tions most frequently used for estimated kidney function III Moderate decrease in GFR 30–59
IV Severe decrease in GFR 15–29
are the Modified Diet in Renal Disease (MDRD) Study V Kidney failure <15 or dialysis
equation [11] for estimating GFR and the Cockcroft-Gault
equation [12] for determining creatinine clearance (CrCl). a
Kidney damage ¼ pathologic abnormalities or markers of disease present
The MDRD equation dominates in epidemiological studies in the urine, blood or on imaging modalities.
[13] and clinical lab use [14–17] when estimating GFR.
The Cockcroft-Gault equation has historically been recom-
mended for use when calculating medication dosages; how- not use the term ‘chronic kidney disease’ but used other terms i.e. ‘patients
with low eGFR’ in reference to their study population were also excluded.
ever, the MDRD equation has recently become acceptable Articles that were deemed potentially appropriate for review based on their
for this purpose as well [18–20]. Choosing which equation abstracts but did not have full-text access via OVID were also excluded.
to use in which circumstance is not clear-cut and both
Data extraction
equations are still criticized for biases and inaccuracy [21,
22]. Newer equations have been proposed to overcome the Abstracts and titles were screened by both authors (J.A. and L.G.G.) and
for those considered potentially relevant full-text versions were obtained to
limitations of the MDRD and Cockcroft-Gault equations. ensure that studies met the selection criteria. The following information
The CKD-EPI (Chronic Kidney Disease Epidemiology was extracted from each article which met inclusion criteria: publication
Collaboration) formula [23], Mayo quadratic equation details, the study’s definition of chronic kidney disease, method used to
[24] and cystatin C-based equations [25–27], respectively, calculate eGFR or CrCl, minimum number of readings (i.e. readings of
serum creatinine) and time between readings necessary to define chronic
claim to be more accurate [23], reliable [28] and supe- kidney disease, method used to determine proteinuria or albuminuria, if
rior [29] to the MDRD and Cockcroft-Gault equations applicable, and method used to determine structural kidney damage, if
especially in patients with higher GFR levels. applicable. The impact factor of each journal included in the review was
As discussed above, there is a lack of agreement within obtained for 2009 [30]. Journals without impact factors were allocated
the literature regarding how to define CKD and estimate a value of <1. Journals that were included in the analysis were also sep-
arated into one of two groups: specialist or general journals. Specialist
GFR despite the clear guidelines from KDOQI. One pre- journals were defined as those dedicated specifically to kidney or urol-
vious study performed a preliminary Google search of the ogy research. No distinction was made between included studies on the
term ‘chronic kidney disease’ and found an absence of any basis of whether CKD was defined for the purposes of selecting a study
consistent definition [2]. It is apparent that variations in the cohort, evaluating CKD as an exposure or studying CKD as an
outcome.
definition of a disease within research will have implications
for further research along with potential implications for Outcomes
practice. To the best of our knowledge, there has not been Each article was assessed for quality as defined by the authors of this paper
a systematic review of literature that describes or quantifies described below. A marker of good article quality was the use of KDOQI
the variation of CKD definitions used in original research CKD definition criteria i.e. the article had to indicate that for the purpose of
their study, CKD was defined using KDOQI criteria or that CKD patients
articles. In light of this, the aim of this study was to provide had kidney function (i.e. GFR, eGFR, CrCl) measured on a minimum of
a description and estimation of the variety of methods used to two occasions separated by a minimum period of 3 months. A marker of
define CKD and measure kidney function in original research poor article quality was the lack of mentioning any CKD definition used
papers that have been published in the year 2009. A secon- for the purpose of the article.
dary objective was to investigate whether the quality of the Statistical analysis
journal, in which the article was published, had any effect on This review was not a meta-analysis. A description of the pooled data
the quality of the methodology used to define CKD. was completed and is summarized in Tables 2 and 3. Quality of the
journal (as represented by impact factor and journal type) and its relation
to article quality (as described above) was analysed by chi-square anal-
Methods yses and are summarized in Table 4. Quartiles were determined for the
impact factor values and journals were grouped accordingly. Journals
Literature search were categorized into four groups according to the value of their impact
An electronic search of MEDLINE through OVID was completed on 12 factor. All statistical test values were two sided, and a P-value of <0.05 was
January 2010. The search term ‘chronic kidney disease’ was used and the considered to indicate statistical significance. Analysis was carried out using
search was further limited to articles in English and those published in 2009. SPSS (15.0).
Inclusion/exclusion criteria
To be included in the review, a study had to be an original research article Results
and include patients with CKD. Articles were excluded from the analysis if
studies reported data from a paediatric population due to the fact that
kidney function is calculated differently from an adult population. Studies Studies identified by the literature search
were also excluded if they reported data solely from a population on
dialysis (peritoneal or haemodialysis) as discrepancies in CKD definition The literature search identified 1158 articles from the initial
are futile at this stage. Review articles, commentaries, case reports and search term and search criteria and Figure 1 shows the
studies with a non-human population were also excluded. Articles that did flowchart of the progress of articles in the review. Upon
a
‘According to KDOQI’ was the only stated information when referring to
the methodology of defining CKD. Kidney function estimation used in original research
b
In articles that used structural damage as a criterion for CKD definition. articles
review of article titles and abstracts, 412 articles were The quantity and duration of time between serum creatinine
deemed potentially relevant based on the aforementioned measurements also varied greatly between research articles
criteria. It was possible to obtain online full-text access for (Table 3). The gold standard of CKD definition as defined by
334 articles. Of the 334 full articles, 301 were included in KDOQI as two eGFR readings <60 mL/min/1.73m2 sepa-
the final review. The final review included articles from rated by 3 months was adhered to in only 20% (n ¼ 60) of the
123 different journals. articles. The majority of articles (52.1%) did not indicate any
further information about serum creatinine measurements
CKD definition used in original research articles used to define CKD for the purposes of their research. Of
the articles that did specify the number of creatinine meas-
There were 26 different methods used to define CKD urements used to define CKD (n ¼ 127), 48% (n ¼ 61) only
(Table 2, panel A). Less than half of the articles (47.8%) used one reading. Estimating GFR or CrCl was done using
defined CKD by using solely an eGFR value. An additional solely the MDRD equation in 62% of the articles (n ¼ 181).
44 articles (14.6%) used eGFR in combination with another The remainder used either the Cockcroft-Gault equation
estimation of kidney damage (i.e. albuminuria, proteinuria (7%) or did not indicate the method used to estimate GFR
or structural damage). A substantial number of articles or CrCl (11.6%).
(n ¼ 26) did not indicate how they defined CKD for the
purposes of their research. Of the articles that used struc-
Article quality versus journal quality or type
tural damage as a CKD definition criterion, one-third
(n ¼ 4) did not specify how they determined this structural The influence of journal characteristics on the quality of
damage (Table 2, panel B). article methodology is shown in Table 4. Chi-square
Journal characteristics
Definitions
Impact factors P-value
Implications for future research and clinical practice 3. Jones CA, McQuillan GM, Kusek JW et al. Serum creatinine levels in
the US population: third National Health and Nutrition Examination
In order to ensure that research is interpretable and able to Survey. Am J Kidney Dis 1998; 32: 992–999
be built on in future research, definitions and terminology 4. National Kidney Foundation. K/DOQI clinical practice guidelines for
used must be consistent. The lack of consistency in defin- chronic kidney disease: evaluation, classification, and stratification.
ing CKD in research articles published in 2009 questions Am J Kidney Dis 2002; 39: S1–S266
5. Glassock RJ, Winearls C. An epidemic of chronic kidney disease: fact
the reliability of conclusions made regarding CKD patients. or fiction? Nephrol Dial Transpl 2008; 23: 1117–1121
Additionally, the majority of research articles are not clar- 6. Go AS, Chertow GM, Fan D et al. Chronic kidney disease and the
ifying the methodology used to define CKD—which may risks of death, cardiovascular events, and hospitalization. N Engl
have implications for future research and possibly clinical J Med 2004; 351: 1296–1305
practice. Many physicians are unsure of how to diagnose 7. O’Hare AM, Bertenthal D, Covinsky KE et al. Mortality risk strat-
CKD in the clinical context [32] and certainly current re- ification in chronic kidney disease: one size for all ages? J Am Soc
search practices are not helping to clarify this uncertainty. Nephrol 2006; 17: 846–853
8. Eriksen BO, Ingebretsen OC. In chronic kidney disease staging the
In addition to those differences identified in this study, use of the chronicity criterion affects prognosis and the rate of
several other distinctions between studies such as those progression. Kidney Int 2007; 72: 1242–1248
relating to construct and predictive validity will need to 9. Coresh J, Astor BC, Greene T et al. Prevalence of chronic kidney
be considered. However, it is important to note that the disease and decreased kidney function in the adult US population:
implications of these differences on internal and external third National Health and Nutrition Examination Survey. Am J Kidney
validity in CKD research remains uncertain. Therefore, Dis 2003; 41: 1–12
recognizing that the current CKD definition and staging 10. Weiner DE, Krassilnikova M, Tighiouart H et al. CKD classifica-
tion based on estimated GFR over three years and subsequent cardiac
may evolve, definitions for CKD should be developed, and mortality outcomes: a cohort study. BMC Nephrol 2009; 10: 26
validated and operationalized for research purposes in a 11. Levey AS, Bosch JP, Lewis JB et al. A more accurate method to
timely manner through international consensus. However, estimate glomerular filtration rate from serum creatinine: a new pre-
once this has taken place, the implications (i.e. selection diction equation. Modification of Diet in Renal Disease Study Group.
bias, misclassification, confounding factors) of defining Ann Intern Med 1999; 130: 461–470
study populations, exposure and outcomes according to a 12. Cockcroft DW, Gault MH. Prediction of creatinine clearance from
serum creatinine. Nephron 1976; 16: 31–41
proposed CKD definition must be rigorously examined.
13. Levey AS, Andreoli SP, DuBose T et al. Chronic kidney disease:
common, harmful, and treatable—World Kidney Day 2007. Clin
Conclusions J Am Soc Nephrol 2007; 2: 401–405
14. Miller WG. Reporting estimated GFR: a laboratory perspective. Am
J Kidney Dis 2008; 52: 645–648
This review of literature found that a variety of definitions are 15. Jones GR, Mathew T, Johnson D et al. Implementation of the routine
being used in original research articles to define CKD. This reporting of eGFR in Australia and New Zealand. Scand J Clin Lab
calls into question the validity and reliability of some CKD Invest Suppl 2008; 241: 23–29
research findings as well as the clinical guidelines that have 16. Lamb EJ. United Kingdom guidelines for chronic kidney disease.
emerged from such research. International consensus in CKD Scand J Clin Lab Invest Suppl 2008; 241: 16–22
definitions is urgently required to rectify this situation and 17. Narva AS. The National Kidney Disease Education Program and
other related efforts in the United States. Scand J Clin Lab Invest
ensure further CKD research is valid, clinically applicable Suppl 2008; 241: 12–15
and generalizable to primary and secondary care. 18. Administration FaD. Guidance for Industry: Pharmacokinetics in
Patients with Impaired Renal Function-Study Design, Data Analysis,
Conflict of Interest Statement. All authors declare no conflict of interest. and Impact on Dosing and Labeling. http://www.fda.gov/downloads/
J.A. and L.G.G. do not have support from any company for their submitted Drugs/GuidanceComplianceRegulatoryInformation/Guidances/
work and do not have any non-financial interests that may be relevant to UCM204959.pdf. (5 July 2010, date last accessed)
the submitted work. The results presented in this paper have not been 19. Hermsen ED, Maiefski M, Florescu MC et al. Comparison of the
published previously either in whole or part. Modification of Diet in Renal Disease and Cockcroft-Gault equations
Author’s Contributions: J.A. (guarantor), L.G.G. both contributed to for dosing antimicrobials. Pharmacotherapy 2009; 29: 649–655
study conception and design. J.A. was responsible for the acquisition of 20. National Institutes of Health NKDEP. Chronic Kidney Disease and
data while L.G.G. and J.A. analysed the data and drafted the article. Both Drug Dosing: Information for Providers. http://www.nkdep.nih.gov/
authors had full access to all of the data in the study and can take respon- professionals/CKD_DrugDosing_508.pdf. (5 July 2010, date last
sibility for the integrity of the data and the accuracy of the data analysis. accessed)
Both authors revised the article and granted final approval to the version
21. Glassock RJ. Estimated glomerular filtration rate: time for a perform-
submitted for publication.
Ethics Approval: Ethics approval was not required for this study. ance review? Kidney Int 2009; 75: 1001–1003
Funding: This study was supported by a bursary from the Western 22. Singh NP, Ingle GK, Saini VK et al. Prevalence of low glomerular
Research and Education Network (WestREN). This bursary was received filtration rate, proteinuria and associated risk factors in North India using
through a peer-reviewed application process from which all authors remain Cockcroft-Gault and Modification of Diet in Renal Disease equation: an
independent and the authors’ work was carried out independently of the observational, cross-sectional study. BMC Nephrol 2009; 10: 4
funding organization. 23. Levey AS, Stevens LA, Schmid CH et al. A new equation to estimate
glomerular filtration rate. Ann Intern Med 2009; 150: 604–612
24. Rule AD, Larson TS, Bergstralh EJ et al. Using serum creatinine
References to estimate glomerular filtration rate: accuracy in good health and in
chronic kidney disease. Ann Intern Med 2004; 141: 929–937
1. Hsu CY, Chertow GM. Chronic renal confusion: insufficiency, fail- 25. Hoek FJ, Kemperman FA, Krediet RT. A comparison between cys-
ure, dysfunction, or disease. Am J Kidney Dis 2000; 36: 415–418 tatin C, plasma creatinine and the Cockcroft and Gault formula for the
2. Ridley J, Baker J, Mills C et al. Chronic kidney disease: what’s in a estimation of glomerular filtration rate. Nephrol Dial Transpl 2003;
name? CANNT J 2009; 19: 19–22 18: 2024–2031