#NephJC Chat
Tuesday Jan 28, 9 pm Eastern
Wednesday Jan 29, 9 pm GMT
Wednesday January 29, 9pm IST
JAMA Intern Med. 2019;
Risk of Nephrogenic Systemic Fibrosis in Patients With Stage 4 or 5 Chronic Kidney Disease Receiving a Group II Gadolinium-Based Contrast AgentA Systematic Review and Meta-analysis.
Sean A. Woolen, Prasad R. Shankar, Joel J. Gagnier, Mark P. MacEachern, Lisa Singer, Matthew S. Davenport.
PMID: 31816007 Full Text at JAMA Internal Medicine
Introduction
We may have contrasting views when it comes to nephrotoxicity associated with iodinated-contrast agents. But when it comes to the use of gadolinium-based agents in advanced CKD Stages IV and V, the nephrology community stands united in saying “NO”, lest nephrogenic systemic fibrosis (NSF) should rear its ugly head.
NSF is described as a scleromyxedema-like process, characterized by woody, indurated skin and sometimes hyperpigmented plaques. Initially recognized by Shawn Cowper in 1998, it quickly became apparent that the fibrotic process is systemic and can involve muscles, myocardium, lungs and pleura, kidneys, and testes. This stimulated the name change from nephrogenic fibrosing dermopathy to nephrogenic systemic fibrosis. Increased collagen thickness, increased hyaluronan, and numerous fibroblasts are seen on biopsy. The disease has a progressive course with contractures, debility, and it usually leads to death.
From 1997 to 2007, more than 500 cases of NSF were reported, and were being added to Dr Cowper’s NSF registry, in an effort to understand the etiology. However, the breakthrough came from Austria in 2006, where Thomas Grobner noticed that 5 of the dialysis patients at their center developed NSF within weeks of undergoing magnetic resonance angiography with gadolinium. This was quickly corroborated by others, and the FDA responded by adding a black-box warning in 2007 that advised avoiding gadolinium-based contrast agents (GBCA) in patient at-risk. The disease largely disappeared after that, though given the long latency (reported to be up to 5.5 years) sporadic cases still appear.
While such recommendations helped eliminate NSF, it meant skipping or delaying contrast in patients in need of a diagnosis. The American College of Radiology (ACR) manual on contrast agents now classifies GBCAs into 3 different groups (Figure 1) and labels the lowest-risk GBCAs as group II agents which have “very low, if any, risk of NSF development”. A key element of this classification is whether exposure to GBCA is confounded or unconfounded. Identifying which exact GBCA may have caused NSF in a particular patient requires that they should have received only one GBCA (unconfounded). In case a patient with NSF has received more than one GBCA in the past, it makes the etiologic link somewhat weaker (hence the name confounded).
Classification of GBCA (reproduced from American College of Radiology Manual on Contrast Media)
The Study
The purpose of this study is to assess the pooled risk of NSF in patients with stage 4 or 5 CKD receiving a group II GBCA.
Methods
Systematic review and Meta-analysis
Followed PRISMA reporting guidelines
Literature from:
Ovid
Embase
Cochrane Central Registry of Controlled Trials
Web of Science
Open Grey
Data bases were searched from inception through January 20, 2019
They excluded: conference abstracts, retracted manuscripts, narrative reviews, editorials, case reports, and manuscripts not reporting the total number of group II GBCA administrations
Yielded 2700 citations using “Stage 4 or 5 CKD with or without dialysis (eGFR <30), administration of group II GBCA (gadobenate dimeglumine, gadobutrol, gadoterate meglumine, or gadoteridol), and incident NSF as an outcome”
After the initial screening, 62 studies were reviewed at manuscript level by 2 blinded team members
16 studies were deemed eligible
Primary Outcome:
Pooled incidence of NSF and associated upper bound of the 95% CI (risk estimate) in patients with stage 4 or 5 CKD receiving a group II GBCA
Risk measure for each of the 4 individual group II GBCAs
Results
Final cohort of 16 studies including 4931 patients (Figure 2); inter-rater agreement was good with a ĸ of 0.68 (CI 0.49-0.87)
Figure 1 from Woolen et al, JAMA IM 2020: Study Flow diagram showing selections of citations
11/16 studies were retrospective
5/16 were prospective
The time frame of investigation across all studies spanned from 1997-2017 and were published from May 2008 through April 2019
7/16 were multi-center studies.
7/16 were US studies
8/16 were European
Follow-up for NSF detection was from 3-72 months in the 15 studies that reported this.
The incidence of NSF in patients with stage 4 or 5 CD across all 16 studies was 0/4931 (0%).
The upper bound of the 2-sided 95% CI (1-sided 97.5% CI) for this pooled estimate was 0.07% (Figure 3)
Figure 2 from Woolen et al JAMA IM 2020 – The 95% CI data stratified by 16 studies, representing NSF incidence across all studies. Pooled refers to pooled exposures of all studies. The wide variation in upper bound CI (0.26%-52.2%) between the studies was due to different sample sizes
Figure 3 from Woolen et al JAMA IM 2020 – The 95% CI data stratified by 16 studies, representing NSF incidence across all studies. Pooled refers to pooled exposures of all studies. The wide variation in upper bound CI (0.26%-52.2%) between the studies was due to different sample sizes
Discussion
Bias:
Because the incidence of NSF across all studies was 0% the authors were not able to perform statistical tests for bias (clustering of outcomes at analysis and adjustment for analysis)
15/16 studies, the assessors of NSF were not blinded to the administration of GBCA
7/16 studies had potential funding bias
Strengths:
In all the 16 studies, there was uniformity in absence of NSF at the start of investigations
There was consistency in administration of GBCA across all groups
Narrow inclusion methods with higher inter-rater agreement, low risk of bias
Limitations:
No universal accepted way to diagnose NSF
Lack of blinding in 94% studies
The review included published cases, and assumes that all NSF cases with class II GBCAs would have been published
From January 2000 to June 2006 there was increased use of GBCA in MRI and a parallel increased in NSF. Grobner et al identified an association between GBCA and NSF. In December of 2006 the FDA issued black-box warning to the use of GBCA in advanced CKD patients. In 2010, the gadolinium warning was updated to specify that only 3 agents namely gadopentetate, gadodiamide and gadoversetamide were responsible for most of the NSF cases. The ACR guidelines state that group II agents can be safely used in patients with advanced CKD and those receiving dialysis (provided they receive dialysis after exposure). Many nephrologists, however, continue to discourage GBCAs for fear of NSF. The current study demonstrates impressive safety with regards to NSF and the use of group II GBCA in CKD patients.
It is impressive that not a single case of NSF was reported in nearly 5000 exposures with group II GBCAs. The upper bound risk of 95% CI was 0.07%. On the other hand, Schieda did report rare cases of unconfounded NSF following exposure to group II GBCA in their review, so the risk is not zero.
Some patients in the current study were on dialysis. These patients are at highest risk of developing NSF. Let’s look at gadolinium and it’s clearance.
Free gadolinium is highly toxic and has been shown to induce fibrosis and tissue necrosis in animals. In the contrast agents, it is often chelated in linear or macrocyclic structure as ionic/charged or non-ionic/non-charged forms. The linear structure and non-ionic binding are more likely to be associated with transmetallation or release of free gadolinium (Figure 5). Gadodiamide is one such example and has traditionally been the most frequently associated with NSF.
Figure 5 - A nonionic linear chelate (gadodiamide), which binds Gd3+ less tightly than other chelates (macrocyclic and ionic), is shown undergoing the process of transmetallation. Other endogenous cations such as Cu2+, Fe3+, Zn2+ and Ca2+ compete with Gd3+ for chelate binding, allowing free Gd3+ to be released into the circulation where it may bind other anions such as phosphate and uremic organic anions. Gd3+, gadolinium; Cu2+, copper; Fe3+, iron; Zn2+, zinc; Ca2+, calcium. Source: Mark A. Perazella; CJASN May 2008, 3 (3) 649-651.
GBCAs are eliminated primarily by the kidneys. In individuals with normal renal function, the half-life of GBCAs is 90 minutes. This can be prolonged to more than 30 hours in patients with advanced CKD. The delayed excretion and prolonged tissue exposure to circulating GBCA is likely a key factor in the relationship between renal insufficiency and NSF. But does dialysis help in eliminating gadolinium and thereby prevent NSF?
In 2006, Saitoh, et. al. showed that 74%, 92%, and 99% of gadodiamide was eliminated by the end of the first, second, and third hemodialysis sessions, respectively. When should time of dialysis occur with relationship to GBCA adminsitration? Though scheduling the MRI just prior to the scheduled dialysis can ensure minimal disruption of the schedule, this is paradoxically when the patient will be most acidemic - and acidemia promotes transmetallation. What about the modality? Whereas a single session of hemodialysis clears 75% of gadodiamide, it takes 22 days of continuous ambulatory peritoneal dialysis to remove 70% (Joffe et al, 1998). However, using 10-15 exchanges per day, Murashima et al (2008) were able to remove 90% of the gadolinium in 2 days. So switching from CAPD to APD/CCPD can indeed enhance the clearance. But doing so also entails more rapid loss of residual kidney function - and residual kidney function, in theory, has a stronger relationship with technique failure from PD and longer term prognosis. Hence any reduction in NSF risk (arguably close to zero in first place from this review) should be weighed against this drawback of increasing PD clearance. Lastly, does enhancing clearance of gadolinium be sufficient? Whats the minimum toxic dose of gadolinium needed to cause NSF?Therefore, one is left wondering whether dialysis can truly prevent nephrogenic systemic fibrosis by clearing gadolinium from blood alone, a question difficult to study or answer, given that the baseline risk itself is so low. These guidelines strike a balance buy suggesting
“In patients who are already receiving dialysis (PD or HD), dialysis should continue after receiving GBCA. HD should be performed the same day as GBCA administration, ideally within 2 to 3 hours of MRI. There is insufficient evidence to support initiation of dialysis, change from PD to HD, or altering dialysis prescription to reduce the risk of NSF.”
Unfortunately, there is no treatment for NSF. However, options include steroids, IVIG, plasmapheresis and extra-corporeal photopheresis. Anecdotal evidence has reported the use of cyclophosphamide, thalidomide, ultraviolet therapy, physical therapy including deep massage technique, pentoxifylline, sodium thiosulphate, alefacept and imatinib for treatment (Khawaja et al 2015 ). Other people advocate higher dialysate bicarbonate concentrations to correct acidosis, longer hemodialysis sessions (5-6 hours) with lower blood flows (250-300 cc/min), but these are purely speculative.
Conclusion
Take home message:The study empowers nephrologists with more data on the safety of group II GBCAs’ use in advanced CKD patients. We could see changing beliefs and practice behavior – contrast enhanced MRI could well emerge as the preferred imaging modality over CE-CT in advanced CKD patients. After all, we have learnt GAD ain’t that BAD.
Summary by Naveen Trehan,
Nephrology Fellow, Detroit, Michigan