Rapidly Progressive Renal Failure in Type 2 Diabetes in the Tropical Environment: A Clinico-Pathological Study

Abstract

Background. Rapid decline of renal function in a diabetic suggests the presence of a nondiabetic kidney disease (NDKD). We designed a prospective study to evaluate the factors associated with a rapid decline in renal function in patients with type 2 diabetes. Methods. Over a 2 and a half year period, all patients with type 2 diabetes who presented with documented doubling of serum creatinine in less than 4 weeks or recently diagnosed advanced renal failure were identified. Patients with prerenal causes, urinary tract obstruction, or systemic disease causing renal failure were not included. Renal histology was studied in all cases. Results. A total of 26 patients satisfied the inclusion criteria. Over 75% had serum creatinine > 4 mg/dL at presentation and 62% were dialysis dependent. Renal histology showed mixed lesions of diabetic nephropathy (DN) and NDKD in 11 cases, only DN in nine, and pure NDKD in six. Diffuse proliferative glomerulonephritis (DPGN) was the commonest NDKD (27% cases), all on a background of DN. History of preceding cutaneous or pharyngeal infection was available in five cases. The proportion of postinfectious glomerulonephritis in diabetics with rapidly progressive renal failure was over six times that of the nondiabetic adult RPRF population during the study period. Four patients had acute interstitial nephritis and three showed crescentic glomerulonephritis. Other lesions included amyloidosis, atheroembolic disease, and renal papillary necrosis (one each). The frequency of microscopic hematuria and retinopathy was similar in those with pure DN and NDKD. Four out of seven cases with DPGN showed partial recovery whereas the other three remained unchanged. Conclusions. About two-thirds of patients with type 2 diabetes presenting with rapid decline of renal function in a tropical environment show NDKD. The high incidence of postinfectious glomerulonephritis in this group is possibly related to the high prevalence of skin and soft tissue infections; and could contribute to progressive kidney disease.

Keywords:

• diabetes mellitus
• rapidly progressive renal failure
• postinfectious glomerulonephritis

Introduction

About 25% of all type 2 diabetics develop progressive kidney disease culminating in end stage renal failure (ESRF).[1] The onset and course of the kidney disease is not as well-defined as in the type 1 disease, but by and large it follows a similar pattern and is amenable to modification by interventions such as tight control of blood sugar and blood pressure and use of angiotensin converting enzyme inhibitors or angiotensin receptor blockers.[2] Progression to ESRF typically takes 5–10 years after clinical disease is discovered.[2&3] Rapid decline of renal function over weeks to months suggests the presence of a nondiabetic cause.[3&4] Quick and accurate identification of such an abnormality is crucial for proper treatment. Several retrospective studies[4-19] have looked at the prevalence of nondiabetic kidney disease (NDKD) in diabetics, but few have evaluated their impact on the clinical course or progression to ESRF.

We designed a study to prospectively evaluate the factors associated with a rapid decline in renal function in patients with type 2 diabetes.

Materials and Methods

This prospective single-center study was conducted over two and a half years. All patients with type 2 diabetes presenting to the Outpatient or Emergency services of the Postgraduate Institute of Medical Education and Research, Chandigarh, a large 1400-bed tertiary care referral center in north India, were screened.

Patients were selected for further evaluation if they showed rapid and persistent elevation in serum creatinine ( × 2 of baseline in less than 4 weeks), or advanced renal failure (serum creatinine > 4 mg/dL) in the absence of previously documented renal disease. Patients who on subsequent investigations were found to have evidence of pre-existing kidney disease, identifiable prerenal factors, urinary tract obstruction, or other systemic illnesses that could result in renal failure were excluded.

The following parameters were recorded for each patient: duration of diabetes, presence of diabetic retinopathy, presence and duration of hypertension and proteinuria, details of drug treatment, including history of nephrotoxic drug intake, and recent infection. All patients underwent percutaneous kidney biopsy using a Biopty gun after an informed consent was obtained. Two patients died in hospital and underwent autopsy. Tissue specimens were examined by light microscopy and immunofluorescence by two independent experienced histopathologists. Other data recorded included complement levels, microbiological studies when necessary, and antinuclear (ANA), antiglomerular basement membrane (GBM) and anti-neutrophil cytoplasmic antibodies (ANCA). Dialysis was given according to standard indications. The NDKD detected on kidney biopsy was treated as per standard protocol.

Patients were followed up until death, return of creatinine to baseline, or referral back to primary care physicians. Outcomes were determined on the basis of renal function at the end of 3 months; and patients were classified into 1) recovered, if their serum creatinine values returned to baseline; 2) improved, if the follow up serum creatinine was within 150% of baseline; or 3) unchanged, if the follow-up serum creatinine was above this value or if the patient remained dialysis dependent.

Results

A total of 238 patients with type 2 diabetes and renal insufficiency were seen during the study period. Out of these, 29 patients qualified for inclusion in the study. Three patients died before a detailed evaluation could be done and, hence, were excluded from further analysis. The mean age of the remaining 26 patients was 46.8 ± 10.3 years, and 21 were males. Diabetes had been diagnosed 6.9 ± 4.5 years (range 1 month–16 years) before presentation and the mean duration of symptoms before hospitalization was 4.9 (range 2–12) weeks. Thirteen patients had stable diabetic nephropathy before the current episode (CKD stage 1: 3 and stage 2: 10), eight had no kidney disease, and no data were available in five.

Anorexia, nausea, vomiting, uremic symptoms, and oliguria were the commonest presenting symptoms (Table 1). Nine patients had fever; and evidence of recent infection was noted in 10 cases. Four patients presented with gross hematuria whereas microscopic hematuria and red cell casts were noted in 16 and four cases respectively. Hypocomplementemia was noted in 12 cases. The serum creatinine values at entry ranged from 3.2–10.8 mg/dL. A total of 16 patients required dialysis.

Table 1. Clinical and laboratory manifestations of rapidly progressive renal failure in patients with type 2 diabetes

Parameter	DN alone (n = 9)	DN + NDKD (n = 11)	NDKD alone (n = 6)	Total (n = 26)
Uremic symptoms	7	10	4	21
Edema	9	10	3	22
Oliguria	6	8	5	19
Fever	1	5	3	9
Hematuria
Gross	0	2	2	4
Microscopic alone	4	8	4	16
Nephrotic range proteinuria	8	7	2	17
Leukocyturia	2	6	4	12
Hypocomplementemia	2	7	3	12
Hypertension	9	11	5	25
Diabetic retinopathy	9	9	2	20
Hemoglobin (g/dL)	9.6 ± 2.8	10.7 ± 2.3	11.8 ± 2.9	10.1 ± 2.6
Serum creatinine (mg/dL)	5.2 ± 2.1	4.4 ± 1.8	4.9 ± 2.0	4.9 ± 1.9

DN: Diabetic nephropathy; NDKD: nondiabetic kidney disease.

Histology revealed diabetic glomerulosclerosis in 20 cases. Eleven of these also showed changes suggestive of co-existing NDKD (diffuse proliferative glomerulonephritis: 7, acute interstitial nephritis: 2, atheroembolic renal disease: 1 and renal papillary necrosis: 1). Six cases showed only NDKD (crescentic glomerulonephritis: 3, acute interstitial nephritis: 2 and amyloidosis: 1).

Diffuse endocapillary proliferative glomerulonephritis (DPGN), characterized by polymorphonuclear infiltration, closure of capillary lumina, and IgG and C3 deposits was the commonest NDKD (27%). In contrast, DPGN was encountered in only 11 out of 281 (3.9%) of nondiabetic adults who presented with RPRF during the study period (p< 0.001). Five of the seven patients with DPGN had evidence of recent infection. This was localized to skin and soft-tissue in four, and the pharynx in the fifth.

Acute interstitial nephritis (AIN) was diagnosed in four cases on the basis of heavy interstitial infiltration with lymphomononuclear cells and eosinophils. Two patients had consumed some unknown medication for minor complaints prior to onset of renal failure. The others denied any such history. Of the three patients with crescentic glomerulonephritis, two had pauci-immune disease whereas the third was of the immune complex type. Serological markers including ANCA (IF and ELISA), anti-GBM antibodies, and ANA were negative in these cases. Amyloidosis was of AA type in the solitary patient, secondary to pulmonary tuberculosis. Two patients died; rapidly progressive renal failure (RPRF) had developed following coronary angiography at another hospital in one and the other had presented with severe urinary tract infection and septic shock. Both underwent autopsies and were found to have atheroembolic disease and renal papillary necrosis, respectively.

Out of the nine cases who showed pure diabetic nephropathy (DN), worsening of kidney function had been preceded by administration of radiocontrast agents in two cases and gastrointestinal fluid losses in one. One patient admitted to regular NSAID intake for chronic headache. Light microscopic lesions of acute tubular necrosis were not seen in any case.

The outcomes of these patients have been summarized in Table 2. Of the seven patients with DPGN, four were given supportive management alone and showed partial recovery of renal function. The other three did not improve spontaneously and, therefore, received oral steroids at 1 mg/kg/day. One of them showed transient improvement but returned to dialysis at the end of 3 months, whereas the other two did not respond. In contrast, 45% of the nondiabetic adults with DPGN seen during the study period recovered completely, whereas another 36% showed improvement in kidney function.

Table 2. Outcomes of rapidly progressive renal failure in patients with type 2 diabetes

Histologic class	Recovered	Improved	Unchanged	Died
DN alone	0	04	05	0
NDKD superimposed on DN
DPGN	0	4	3	0
Acute interstitial nephritis	1	1	0	0
Atheroembolic disease	0	0	0	1
Renal papillary necrosis	0	0	0	1
NDKD alone
Acute interstitial nephritis	2	0	0	0
Crescentic glomerulonephritis	1	1	1	0
Amyloidosis	0	0	1	0

DN: Diabetic nephropathy; NDKD: non-diabetic kidney disease; DPGN: diffuse proliferative glomerulonephritis.

^*Kidney function remained unchanged in all cases who died.

Crescentic glomerulonephritis was treated with three doses of intravenous methylprednisolone (1 g/d) followed by oral prednisolone at 1 mg/kg/d. The two patients with pauci-immune disease also received oral cyclophosphamide at 2 mg/kg/day. Renal function returned to normal in one and the other two showed partial recovery. All four patients with AIN received steroids. Three recovered; the fourth, who had coexisting diabetic glomerulosclerosis, showed partial improvement, but progressed to ESRF over one and a half years. Of the nine patients with lesions of pure DN on histology, four cases improved on supportive management alone, whereas the rest were unchanged.

Discussion

Diabetic nephropathy is characterized by the appearance of proteinuria, hypertension, and slowly progressive renal failure leading to ESRF over 8–10 years in a person with diabetes.[1-3] Kidney biopsy is not usually performed to confirm the diagnosis. Diabetic nephropathy is the commonest cause of ESRD in the west, and is being encountered increasingly frequently in the developing countries. The contribution of NDKD to the kidney disease burden in diabetics, especially those with type 2 disease, remains a matter of speculation. It is generally accepted that once detected, the progression of DN is predictable. Rapid decline of renal function over a period of weeks is not a feature of DN and should trigger a search for other causes.

In the two and a half year study period, about 11% of all patients with type 2 DM and renal failure presented with RPRF incompatible with the known natural history of DN. This figure, however, is possibly an overestimate since patients with this kind of presentation are more likely to be referred to a tertiary care center.

About 65% of patients presenting with rapid decline of renal function showed NDKD in this study. The striking finding was the high frequency of DPGN. About 27% of all cases and 41% with NDKD exhibited this lesion. This frequency was far in excess of the proportion of DPGN seen in nondiabetic adults with RPRF at our center during this period. The reported incidence of all forms of glomerulonephritis in type 2 diabetes is reported at 0%–67%. Postinfectious glomerulonephritis (PIGN) forms 0%–17% of all NDKD (Table 3). There are several possible reasons for this variation. Most of the older studies have been retrospective, and criteria for selecting cases for biopsy are unknown. Many studies have used absence of retinopathy as the major indication for biopsy. Rapid deterioration of renal function was the reason for biopsy in 19 out of 32 cases reported by Nzerue et al.[16] Pinel et al.[18] excluded all patients with clinical renal disease other than presence of type 2 diabetes and albuminuria, and did not report any complicating renal disease. The study by Fioretto et al.[20] on micro-albuminuric patients also did not find increased incidence of nondiabetic renal disease. Waldherr et al.[4] found NDKD in only one out of 210 consecutive autopsies in diabetics over a 2-year period.

Table 3. The frequency of diabetic and nondiabetic kidney disease in various studies

	Reference Number
	4	7	8	9	10	11	12	13	14	15	16	17	19	20	Present study
Number of cases	210	60	128	35	35	52	80	32	88	51	32	49	33	46	26
Diabetic glomerulopathy
Nodular	37	47	25	8				22	38	34	10	1	20	24	16
Diffuse	129		83	20							17	36	12		4
None	44	13	20	7				12						22
NDKD Glomerulonephritis	1	16	20	8	11	16	53	15	36	17	18	6	3	8	10
DPGN		2				6	14	3	5		7				7
MGN		3	7		5	2	6	3	8					3
IgA nephropathy	1	2	13		1	2	6			10			1	1
Focal proliferative					5	2	5	1			6			1
Crescentic		4				2	5	2						1	3
Mesangial proliferative				3			4	3	17				1
Membranoproliferative		1						1	4	1				1
Lupus nephritis		1					1							1
Minimal change		2		4		2	12	2	2	1	2
Others		1		1									1
Others		1	0	0	3	1	10	17	0	4	0	0	1	9	7
Amyloidosis					1	1	2							3	1
Myeloma kidney														2
Acute cortical necrosis														1
Interstitial nephritis					1		8	17					1		4
Papillary necrosis														2	1
Atheroembolic disease														1	1
Hypertensive nephrosclerosis		1								4
Sarcoidosis					1

Infection was documented in five out of seven cases with DPGN, a majority being neglected soft-tissue infections of the lower limbs. The cause-and-effect relationship was based on a temporal relationship between the infections and biopsy. This reflects the poor standard of hygiene, lack of foot care, and limited access to healthcare amongst the economically disadvantaged diabetic population living in rural areas of tropical countries. John et al.[12] noted PIGN in 17.5% of all patients with NDKD presenting to a south Indian hospital. Rarely PIGN is reported now by Western hospitals. An exception was the study of Nzerue et al.,[16] who found immune complex glomerulonephritis in seven of their 32 African-American type 2 diabetics, most of whom came from inner-city populations and had relatively poor access to medical care. A postinfectious etiology could be established in only two of their cases.

Interestingly, DPGN was superimposed on diabetic glomerulosclerosis in all our cases. Similarly, six out of seven cases with immune complex GN in another series had coexistent diabetic nephropathy.[16] It has been suggested that diabetics may be predisposed to glomerular immune complex deposition, possibly because of hyperfiltration, altered mesangial function, and preexisting alterations in basement membrane. Larger and more numerous immune complexes were seen in diabetic rats upon induction of Heymann nephritis than their nondiabetic counterparts.[21]

Our patients with DPGN on a background of DN showed significantly worse outcomes compared to their nondiabetic counterparts; 85% required dialysis and the renal function did not return to baseline in any case. This was in sharp contrast to the nondiabetic adults with PIGN, who showed complete and partial recovery in 45% and 36% of cases respectively. It has been speculated that PIGN could accelerate progression to renal failure in diabetics. Chihara et al.[14] found that diabetics with glomerulonephritis progressed faster to ESRF, especially where the complicating lesion was mesangiocapillary glomerulonephritis.

Diseases involving other compartments of renal parenchyma were also encountered, with AIN being the commonest. Some degree of interstitial lymphomononuclear infiltration is seen in diabetic nephropathy.[22&23] However, we found heavy infiltrate consistent with a diagnosis of AIN in 16% of the cases. Half the patients admitted to intake of some drug just before the development of AIN. It is a common practice in rural areas of the developing world to get medications from local chemists or practitioners of indigenous systems of medicine, and records are usually not available. Improvement following steroid therapy corroborated the diagnosis of AIN. In contrast to glomerulonephritis, 75% of all patients with AIN recovered.

No clear cause of rapid decline in kidney function could be identified in about 35% of cases, and kidney biopsy showed only diabetic nephropathy. The clinical presentation of these cases was similar to the other cases. Prior records of renal function tests were not available in about 19% of our patients. Hence it is possible that at least some of them had preexisting but clinically silent renal insufficiency, and the symptomatology was unmasked by minor clinically unidentifiable insult.

Presence or absence of retinopathy and/or hematuria was not helpful in differentiating NDKD from DN alone in the present study. Some studies[12],[14] have found absence of retinopathy to be useful in identifying patients with NDKD. Kasinath et al.[5] and Amoah et al.[7] found low incidence of retinopathy in patients with diabetic glomerulosclerosis due to type 2 disease. Retinopathy can be missed during a simple bedside ophthalmoscopy performed by a medical resident. All our cases were examined by trained ophthalmologists.

In conclusion, compared to the nondiabetic population with RPRF, DPGN is six times more common in type 2 diabetics who present with RPRF in tropical countries. This is possibly related to the high prevalence of skin and soft tissue infections due to diabetes, coupled with poor access to healthcare. Compared to uncomplicated DPGN, the prognosis is worse when this lesion is superimposed on diabetic glomerulosclerosis.