Should Insulin-dependent Diabetic Patients be Screened for Malnutrition Before Total Joint Arthroplasty?

A Cohort at Risk

Andrew M. Schneider, MD; Nicholas M. Brown, MD


J Am Acad Orthop Surg. 2021;29(15):673-680. 

In This Article


The correlation of malnutrition with morbid obesity is generally well accepted, and sufficient evidence exsist to support the utility of preoperative serum albumin screening before TJA in patients with BMI ≥40 to minimize postoperative complication risk.[12,13] However, despite its association with poor nutrition, the diabetic cohort undergoing TJA is less scrutinized for malnutrition; the current literature gives little insight into preoperative screening recommendations and risk stratification of patients within this demographic. With the increasing incidence of diabetic patients undergoing TJA, characterizing the risk of malnutrition and its association with postoperative complications in this cohort and its subgroups would be of value.[22] Our goal was to compare the risk of malnutrition between diabetic patients and morbidly obese patients undergoing TJA. In addition, we sought to investigate the association of malnutrition on postoperative TJA complication and infection risk in the diabetic cohort.

In the current study, we identified a cohort of 203,277 TJA patients from 2011 to 2018 and found that those patients with IDDM were at a particularly high risk of being malnourished in both the morbidly obese and nonmorbidly obese populations (frequency of 11.9% and 9.9%, respectively). Furthermore, patients with IDDM without morbid obesity had a 1.5x higher frequency of malnutrition than morbidly obese patients without diabetes (9.9% versus 6.4%, respectively; P < 0.001). Among patients with co-occurring diabetes and malnutrition, patients with IDDM were at greatest risk for 30-day postoperative complications and infection (OR 2.081 [1.652, 2.621]; P < 0.001 and OR 1.894 [1.231, 2.913]; P = 0.004, respectively).

Our findings are supported by the current literature. In a retrospective review of the NSQIP database involving 114,102 TKA patients, Webb et al[23] sought to determine the postoperative adverse event rate among diabetic subpopulations. Controlling for potential confounders, patients with IDDM were found to be at notably greater risk of 12 postoperative complications at 30 days compared with patients without diabetes, although patients with NIDDM were only found to be at greater risk for two postoperative complications (myocardial infarction and length of stay) compared with patients without diabetes. The authors advocated for a heightened awareness of insulin dependence as a predictor of postoperative adverse outcomes in TKA patients. In a retrospective single institution review of primary TKA over a 17-year period (1995 to 2011), Watts et al[24] investigated TKA outcomes among diabetic subpopulations in the morbidly obese (BMI ≥ 40) cohort specifically. At the 6-year follow-up, they found no difference between patients without NIDDM and diabetes. However, patients with IDDM had an increased risk of reoperation, revision, and deep infection as compared to patients without diabetes, leading the authors to encourage accounting for diabetes severity when risk stratifying patients before surgery. Lovecchio et al[1] retrospectively examined 43,299 TJA patients and similarly stratified them into NIDDM and IDDM groups. Although both NIDDM and IDDM patients were found to be at increased risk for the development of postoperative complications, only patients with IDDM had increased risk for 30-day mortality and readmission. Regarding malnutrition and TJA outcomes, Courtney et al[12] reported that morbidly obese patients undergoing TJA were more likely to be malnourished than nonmorbidly obese patients (19% versus 11%, respectively; P = 0.010). In addition, they found malnutrition to be an independent risk factor for postoperative complications, corroborating many previous studies.[6,10,25] They suggested that preoperative serum albumin screenings for malnutrition in the morbidly obese cohort can help identify this particularly at-risk cohort so that nutritional intervention could take place before surgery.

The aforementioned studies have importantly identified IDDM and malnutrition as independent risk factors for complications after TJA, corroborating the results of the current study. Moreover, malnutrition has been identified as prevalent in the morbidly obese cohort and has served as a target for screening and optimization before TJA. However, to our knowledge, no previous report has evaluated the risk of malnutrition among patients with diabetes undergoing TJA and compared it with the risk of malnutrition in the morbidly obese cohort. In addition, the additive complication risk of both malnutrition and diabetes, stratified by diabetic subgroup, has yet to be elucidated in the recent literature.

The explanation for the high frequency of malnutrition in the diabetic cohort of our study, and among the insulin-dependent patients in particular, is likely multifactorial. In Western nations, low-income populations are more likely to develop diabetes.[26] Disadvantaged socioeconomic status may lead to poor nutritional habits, limited knowledge of proper dieting techniques, and hazardous concomitant conditions such as drug or alcohol abuse.[26] Having diabetes, either IDDM or NIDDM, is a risk factor for developing depression and having depression is a risk factor for developing diabetes.[27] Depression may instigate poor nutritional behaviors, such as binge eating or anorexia, both of which can negatively affect nutritional status.[28,29] Biologically, the chronic inflammatory state generated by diabetes can lead to altered bowel conditions characterized by malnutrition, such as gastroparesis and celiac disease.[14–16]

Basic science evidence also exists to support the close link between insulin and albumin. Albumin is an essential biomarker used to assess liver function, and multiple factors, including nutritional state, regulate its production.[30–33] In diabetes, the serum concentration of albumin is decreased, requiring the administration of insulin to prevent hypoalbuminemia.[34–36] Biochemical studies have elucidated that insulin stimulates albumin production by activating gene transcription.[31,34,36–39] Chen et al[35] demonstrated insulin and albumin's intimate relationship by showing that insulin acts directly on the liver through the IR/PI3K/Akt pathway to inhibit Foxo1, which functions as a repressor of albumin expression.

In 2019, Schroer et al[9] conducted a prospective analysis demonstrating that malnutrition, represented by hypoalbuminemia, was a modifiable risk factor before TJA. In a study of over 4,700 patients, they found that patients with a preoperative serum albumin level ≤3.4 who underwent a nutrition intervention program involving a high protein, anti-inflammatory diet had shorter length of hospital stay, lower primary hospitalization charges, lower charges associated with hospital readmissions, and lower 90-day total charges compared with patients with hypoalbuminemia who did not have nutritional intervention before TJA (P < 0.001). This study lends credence to the idea that hypoalbuminemia is a reliable marker for malnutrition, and malnutrition is a modifiable risk factor for adverse outcomes in the TJA cohort.

It is crucial to understand the results of this study in the context of its limitations. First, although this current study's findings support the notion of preoperative albumin testing in the IDDM cohort, the economic impact of this screening, which would require a comprehensive cost-benefit analysis, is outside the scope of the study. Second, our data are limited to that which is provided by the NSQIP database. Because of the large percentage of TJA patients in the database without recorded albumin (48.1%), we did two additional analyses to better understand the potential bias this introduces in our results. The first additional analysis compared the cohort with missing albumin with the cohort with recorded albumin levels to assess for differences. Our findings show that the two groups are similar regarding age and sex, and patients with missing albumin smoked less and had lower ASA scores (Table, Supplemental Digital Content 3, The second additional analysis compared the infection rate and overall complication rate between patients with albumin <3.5, patients with albumin ≥3.5, and those with no albumin recorded. The complication and infection rates of those with no recorded albumin were nearly identical to those with albumin ≥3.5, and both groups had lower complication and infection rates than the cohort with albumin<3.5 (Table, Supplemental Digital Content 4, Therefore, it is possible that our data actually underestimates the effect of malnutrition and IDDM on complications. Third, the designation of IDDM versus NIDDM in the NSQIP database does not provide any insight into disease control. Recent data exist to suggest that variables such as perioperative hyperglycemia and glucose variability adversely affect TJA outcomes, neither of which are captured in the NSQIP data set.[40,41] For instance, the TJA infection and complication risk of a well-controlled insulin-dependent diabetic compared with a poorly controlled noninsulin dependent diabetic patient are unable to be evaluated by the current study, and thus, our conclusions should be cautiously interpreted in this context. Fourth, outcome data are limited to 30 days postoperatively in the NSQIP, which may affect some findings more than others. For instance, complications such as deep infection or reoperation would likely be greater if follow-up was longer. Fifth, we acknowledge that biomarkers other than albumin, such as transferrin, prealbumin, and total lymphocyte count have also been described to evaluate for malnutrition. However, given the availability of preoperative albumin in the NSQIP, its widespread clinical use, and its use as a malnutrition marker in similar studies to ours, we believe that using it as a surrogate for malnutrition in this study was appropriate. Finally, patients with diabetes were stratified based on their use of insulin or oral medications. It is possible that patients with diabetes or prediabetes were included in the nondiabetic group if they were not using any medications to properly control their hyperglycemia.

In conclusion, the current study found that patients with IDDM were at a particularly high risk of being malnourished in both the morbidly obese and nonmorbidly obese populations. In addition, patients with IDDM but without morbid obesity were 1.5x more often malnourished than morbidly obese patients without DM, a cohort that is routinely screened for malnutrition before TJA. Among patients with co-occurring diabetes and malnutrition, patients with IDDM were at greatest risk for 30-day postoperative complications and infection. Given these findings, it is imperative to have a heightened suspicion for malnutrition among patients with IDDM undergoing TJA, and preoperative malnutrition screening should be considered in this cohort. Our results show that the unique combination of IDDM and malnutrition puts patients at particularly high risk for postoperative complications and infection; detection of malnutrition preoperatively can serve as an area for targeted intervention to minimize the risk of adverse outcomes after TJA.