J Cardiovasc Med. 2017 Aug;18(8):625-630.

Altered fasting glycemia in cardiac patients during in-hospital rehabilitation: impact on short and long-term follow-up.

Masnaghetti SE1, Sarzi Braga S, Vaninetti R, Baiardi P, Pedretti RFE.

Department of Cardiology, Fondazione Salvatore Maugeri I.R.C.C.S, Tradate, Varese bDepartment of Statistics, Fondazione Salvatore Maugeri I.R.C.C.S, Pavia, Italy.



Hospitalized patients after acute cardiovascular events have poorer prognosis if glucose regulation is diagnosed as abnormal. We compared the short and long-term outcome of patients with newly diagnosed altered fasting glycemia (AFG) to that of known diabetic patients and patients with normal glucose regulation (NGR) after admission to cardiac rehabilitation.


We retrospectively analyzed 2490 consecutive patients. Three groups were identified: known diabetes mellitus (n = 540, 22%), fasting glycemia above 110 mg/dl (AFG, n = 269, 11%), and fasting glycemia 110 mg/dl or less (NGR, n = 1681, 67%). Clinical variables, complications, and all-cause mortality were evaluated.


At follow-up (median 3.1 ± 2.4 years), after adjustment for age, sex, BMI, left ventricular ejection fraction, history of coronary artery disease, AFG had a significantly longer hospital stay versus NGR (21 ± 8 versus 20 ± 8 days; P = 0.019) and higher risk of paroxysmal atrial fibrillation (P = 0.041), pleural/pericardial effusions (P < 0.001), skin complications (P = 0.033), other events (P = 0.001), and blood tests (urea: P = 0.007; white blood cells: P = 0.002; neutrophils: P < 0.001; creatinine: P = 0.022). All-cause mortality was significantly higher in diabetes mellitus versus NGR (odds ratio 1.61, 95% confidence interval 1.17-2.21); a nonsignificant trend was observed in AFG versus NGR (odds ratio 1.23, 95% confidence interval 0.77-1.98).


A high AFG prevalence in cardiac patients admitted to rehabilitation was observed. AFG patients were more vulnerable than NGR patients, had higher complication rates independently of covariates, and required longer hospital stay. AFG was not a significant predictor of all-cause mortality at 3 years, whereas DM was.




At admission to a in-hospital cardiac rehabilitation programme we observed that many not diabetic patients had a high fasting glycemia value. An association between this finding and poorer prognosis during the acute phase of myocardial infarction was known but we could not find information about it’s possible significance during the subacute – chronic phase of rehabilitation.

We retrospectively analysed 2490 patients accepted after recent congestive heart failure (227 – 9%), coronary artery disease/percutaneous coronary intervention (628 – 27%), cardiac surgery (1428 – 57%), other diagnosis  (157 – 6%) and identified 3 groups of patients according to history of diabetes or to the single fasting glycemia value at admission: 540 (22%) patients had a history of diabetes (DM); 1,681 (67%) patients had normal glucose regulation (NGR) and 269 (11%) had altered fasting glycemia (AFG).

AFG and DM patients were significantly older (p<0,001), had a higher BMI (p<0,001 and were more frequently hypertensive (p<0,001) than NGR patients. No differences were observed in gender and diagnosis at admission.

Statistical analysis was performed on length of hospital stay, blood exams and clinical events during the rehabilitation period. Death was considered for the analysis both during hospitalization and at follow up. The results of the analysis were adjusted for age, gender, BMI, history of hypertension, LVEF, and diagnosis at admission (coronary artery diseases/percutaneous coronary intervention, cardiac surgery, congestive heart failure).

In the AFG compared to NGR group of patients a longer length of hospital stay (21.2±9.2 vs 19.6±8.9 days; p=0.019), a higher level of white blood cells count (p=0,002) and hemoglobin value (p<0,001), a higher risk of paroxysmal atrial fibrillation (p=0,041) and cutaneous complications during hospitalization (p=0,033) were observed while in both AFG and DM patients urea (p=0,007 & p<0,01), neutrophil (p<0,001 & p=0,028) and creatinine values (p=0,022 & p<0,001), pleural/pericardial effusion (p<0,001 both) and other significant events (p=0,001 & p=0,048) were significantly higher.

In DM patients, there was a significantly higher risk of renal failure compared to NGR patients (p<0,001), while a trend to higher incidence, but not statistically significant, was present in AFG vs. NGR patients (p=0.06).

At the analysis of mortality at a mean follow up of 3.1 ± 1.4 years a 9% all-cause cumulative mortality was registered in the population: 81 in DM (15%), 31 in AFG (11.5%) and 118 patients in the NRG group (7%) died, respectively.

At longitudinal analysis were excluded 10 patients who died during the rehabilitation programme and 79 (15%), 28 (10%)and 113 (7%) deaths were registered in DM, AFG and NGR groups, respectively. The Cox model showed a significantly higher risk of mortality in both DM and AFG compared to NGR but after adjustment for risk factors  only the DM compared to the NGR group revealed a higher risk of mortality (HR 1.61; CI 1.17–2.21; p=0.003).

We conclude that although a single value of high fasting glycemia could be considered insufficient for diagnosis of altered glucose metabolism, according to diabetes guidelines, it was able to select a higher risk population after all. The retrospective nature of the study limited the selection of the patients and a possible in-homogeneity in AFG group should be considered. We believe that a significant association between AFG and death might be possible at a long lasting follow up.

The message we can take from this study is that the best treatment of the patients depends on our knowledge of the illness and a high value of glycemia should be taken into account.