5.3 Phenacetin kidney

Definition: Renal disease due to phenacetin containing mixed analgesics

“Phenacetin kidney”

(“Analgesic nephropathy” due to phenacetin containing mixed analgesics)

In 1953, Spühler and Zollinger described “Chronic interstitial nephritis”.

In this publication, the “Phenacetin kidney” with papillary necrosis was described in all details. However, drugs (not mentioning phenacetin) were only discussed among other aetiological factors). In 1955, Zollinger described „Chronic interstitial nephrits in patients with an excessive intake of phenacetin-containing analgesics (Saridon ®). This syndrome - characterized by papillary necrosis, chronic interstitial nephritis, cortical atrophy and brown pigmentation of the papillae and urothelium – was later termed “Analgesic nephropathy”.

Removal of phenacetin from pain killers – around 50 years after the first description in Switzerland – brought the solution overnight: the “Phenacetin kidney”, or as then falsely labelled “Analgesic nephropathy“ disappeared from the daily clinical routine despite the fact that mixed analgesics have continued to be popular drugs, with paracetamol replacing phenacetin in many preparations since the early 1980s.  Today medical students, nephrologists and even pathologists are no longer aware of this disease.

Currently, “Analgesic nephropathy” may be used by clinicians to label renal dysfunction in patients with a documented abuse of any kind of analgesics, but from the morphological point of view “Phenacetin kidney” is a morphologically specific variant of “Analgesic nephropathy”. Whether non-steroidal anti-inflammatory drugs cause similar or identical morphologic lesions as described below is as yet unknown.

Thus, we suggest the following nomenclarure:

                        Analgesic nephropathy

                        a, due to phenacetin containing drugs

                        b, due to nonsteroidal anti-inflammatory drugs (different from  acute interstitial nephritis),

                        c, due to other drugs

Pathology of “Analgesic nephropathy” due to phenacetin containing drugs

(so-called “Phenacetin kidney”)

Classical  “Phenacetin kidney” as described below is induced by mixed analgesics containing phenacetin, acetylsalicilic acid and coffeine or similar combinations. It is a special and the only well-characterized variant of a whole spectrum of renal diseases which may be induced by analgesics e.g. non-steroidal anti-inflammatory drugs and others.

The development of the classical form of “Phenacetin kidney” from the earliest detectable lesion in the urinary tract to the terminal stage of bilaterally contracted kidneys will be traced below.

Three stages can be identified:

            • capillary sclerosis of the lower urinary tract

            • papillary necrosis

            • cortical changes

Capillary sclerosis

The earliest detectable change in phenacetin abuse is capillary sclerosis of the urinary mucosa, found in 80-90 per cent of cases. Basement membrane thickening is seen in the capillaries immediately underlying the urothelium, which itself is unaltered and in severe cases also in the muscle layer. With increasing severity the basement membrane thickens, with constriction of the lumen and eventually complete occlusion. Basement membrane thickening has been shown by electronmicroscopy to result from the concentric deposition of thin basement membrane lamellae, in an onion or tree ring structure. Massively constricted or occluded capillaries may exhibit over twenty such lamellae, entrapping polymorphous cell debris and a dense concentration of lipids, visible on Sudan staining. In particularly severe cases the lipid deposition causes a macroscopically visible brownish pigmentation of the urinary tract mucosa. Such hyperpigmentation is a particular feature of the abuse of phenacetin-containing analgesics, and it is not described in abuse of phenacetin-free analgesics. Capillary sclerosis increases in severity from the calices to the pelvi-ureteric junction, is most marked in the proximal ureter and gradually decreases to the bladder where in particularly severe cases, it may be found, especially around the ureteric ostia.

In mild capillary sclerosis, evidence of nephropathy is absent in some half of patients. In moderate or severe cases, nephropathy is still undetectable in 15 per cent. On pathological examination, papillary necrosis is detectable macroscopically in 15 per cent of patients with mild capillary sclerosis and in 40 per cent of those with moderate and severe disease.

Development of papillary necrosis

The morphology of papillary necrosis is well-documented since the fundamental studies by Burry and Gloor  and can be divided into three stages: early, intermediate, and late, or full-blown papillary necrosis.

In the early stage of papillary necrosis, the macroscopic appearance of the kidney is virtually normal; however, in abuse of phenacetin-containing analgesics, the medullary pyramids show yellow-brown streaks converging on the papillary tips.

Histologically, the lesion is confined to the central part of the inner medulla. The outer medulla and cortex are unaffected. Transverse sections through the papillary tips show patchy lesions distributed around the collecting duct bundles. The interstitium is widened by an increase in ground substance and contains markedly altered peritubular capillaries and ascending limbs of Henle's loops, adjacent descending limbs, collecting ducts and vasa recta show little or no change. The basement membranes of the ascending limbs and peritubular capillaries are thickened by layers of newly formed basement membrane lamellae. Thus the earliest detectable changes in papillary necrosis are: peritubular capillary sclerosis and analogous changes in the basement membrane of the ascending limbs of the loops of Henle (tubular sclerosis). The tubular epithelium of the ascending limbs and the peritubular capillary endothelial cells show degenerative changes or are fully necrotic. Interstitial cells likewise die. Sudan-stained sections from abusers of phenacetin-containing analgesics show lipid infiltration of the thickened basement membranes and surrounding interstitium. Focal deposits of calcium salts in the basement membranes are seen also.

In the intermediate stage of papillary necrosis, foci become confluent in the papillae and extend up to the border of the outer medulla. The lateral parts of the inner medulla, which receive the collecting ducts of the columns of Bertin, remain well preserved, as do the collecting ducts in general and the vasa recta.

In late-stage, full-blown necrosis, the collecting ducts and vasa recta also are destroyed. Clefts develop within the necrotic areas, especially bordering viable renal parenchyma and, among other features, become lined with collecting duct epithelium. Papillary necrosis is a non-reactive, granulocyte-free necrosis. Granulocytes also are virtually absent from adjacent viable tissue. Lymphocytes and histiocytes are common in the area surrounding the vascular bundles in the outer medulla.

Outcome of papillary necrosis and cortical change

Papillary necrosis gives rise to a variety of scenarios: rapid demarcation and sloughing of necrotic papillae, protracted sloughing, and persistence of necrotic papillae in situ.

If a necrotic papilla sloughs partially or totally, the papillary cavity can reepithelialize. Necrotic papillae remaining in situ can calcify and later ossify. For a long time a ghost outline of the original structure can still be detected in the necrotic papilla, as shown by histological examination of papillary fragments spontaneously voided in urine.

What happens to the necrotic papillae is decisive for the renal cortex. Thus if papillary sloughing is rapid, cortical changes may be absent. However, this rarely occurs.

If sloughing is delayed or incomplete, and particularly if the necrotic papillae remain in situ, progressive cortical atrophy ensues, with interstitial fibrosis and streaky infiltration by lymphocytes and histiocytes (chronic, non-destructive, interstitial nephritis). If some necrotic papillae rapidly slough, while others remain in situ, interstitial nephritis and cortical atrophy develop only with respect to the non-separated papillae. The conclusion to be drawn from this relationship between papillary necrosis on the one hand, and interstitial nephritis on the other, is that chronic interstitial nephritis is the consequence, not the cause, of papillary necrosis. Obstruction of the intrarenal urine flow due to papillary necrosis leads to chronic interstitial nephritis, as does urinary obstruction from other causes.

Since papillary necrosis affects primarily the central papillary area, and spares the lateral area until much later, hypertrophy can occur of the cortical tissue of the columns of Bertin, draining urine mainly via the collecting ducts that open laterally into the papillae. Compensatory hypertrophy of the columns of Bertin offsets the suprapapillary loss of parenchyma for a long while, and explains why glomeruli showing focal segmental glomerulosclerosis and hyalinosis (Zollinger's so-called overload glomerulitis) are a particularly common finding in this hypertrophied area of cortex . In the atrophic suprapapillary cortex, however, non-specific glomerular changes predominate, with periglomerular fibrosis, collapse or global sclerosis.

Full-blown analgesic nephropathy

Full-blown analgesic nephropathy is the result of a complex series of events in which primary analgesic-induced lesions overlap with secondary lesions. The kidneys are bilaterally contracted and in part ridged, with the contracted areas corresponding to papillary necrosis and the ridges to the hypertrophied columns of Bertin. On the cut surface the corticomedullary junction is poorly demarcated. Some papillae have sloughed, while others remain in situ. The renal pelvis is of normal or slightly increased width and may contain sloughed fragments of necrotic papillae. Brown pigmentation of the urinary mucosa may be present. Papillary necrosis becomes difficult to identify in more severely contracted kidneys. Horizontal sections no longer suffice, and transverse sections across the calices are required. In the terminal stage, even the previously hypertrophic tissue of the columns of Bertin becomes atrophic, resulting in small smooth kidneys.

Until the intermediate stage of papillary necrosis, the predominant lesion is capillary sclerosis, with analogous basement membrane changes in the loops of Henle. With the development of full-blown papillary necrosis, cortical changes are induced which result in suprapapillary cortical contraction and simultaneous compensatory hypertrophy in the relatively spared columns of Bertin that remain.

Complications of analgesic nephropathy

Complications frequently overlay and blur the typical picture of analgesic nephropathy and occasionally prevent the correct diagnosis from being made. There are essentially four complications: hypertension, pyelonephritis, hydronephrosis or pyonephrosis, urolithiasis, and urothelial tumours.

There is a direct correlation between frequency of hypertension and severity of renal failure. Thus hypertension is observed in approximately 80 per cent of patients with a serum creatinine of 885 µmol/l, but in only 30 per cent with a serum creatinine of 177 µmol/l. Additional lesions are found in the renal cortex in hypertension: arteriosclerosis, arteriolosclerosis (rarely), hypertensive glomerulosclerosis and renovascular scarring. Renal contraction is aggravated by renovascular scarring. Hypertension-induced lesions often are more readily recognized in the columns of Bertin than in the markedly contracted supra-papillary cortical tissue. A rare complication is the central arterial contracted kidney, which may occasionally be bilateral, due to arteriosclerotic stenosis of the renal artery.

Acute and chronic pyelonephritis occur in approximately 75 per cent of cases of analgesic nephropathy. The necrotic renal papillae are an ideal culture medium for bacterial infection. But it is open to question whether scarring in interstitial destructive nephritis is due to bacterial infection. Acute pyelonephritis with papillary necrosis is a red herring, which can result in failure to recognize analgesic-induced papillary necrosis.

Similarly the typical picture can be completely obscured by hydronephrosis and pyonephrosis from blockage of the lower urinary tract by sloughed necrotic papillae. A correct diagnosis is only reached in advanced renal destruction, particularly in nephrectomy specimens, if capillary sclerosis is deliberately looked for or if typical analgesic-induced sloughed papillae can be identified in the ureteral lumen causing the obstruction. Capillary sclerosis can occasionally be destroyed by necrotizing mucosal inflammation, making it impossible to diagnose analgesic abuse on morphological grounds alone.

Secondary cyst formation after long term hemodialysis may also obscure the typical picture of analgesic nephropathy. As soon as the whole kidney is permeated by cysts of variable size, the necrotic papillae are easily overlooked.

Cessation of analgesic abuse

There have been few clinical and morphological investigations into what happens after analgesic abuse is discontinued. However, analysis of individual cases suggests that capillary sclerosis halts. Partial regression of the basement membrane changes can occur, through the continuous turnover - build up and breakdown - of the basement membrane. Ureters which over the years have been excluded from urine flow show no further capillary sclerosis. Remains of basement membrane in the stroma can be detected only by electron microscopy. If the assumption is correct that capillary sclerosis regresses, this would explain why no further capillary sclerosis is found in patients dying ten or more years after cessation of analgesic addiction. The hypothesis would also explain why renal function remains stable, or even improves, when analgesic abuse is abandoned in the early stages. This favourable outcome is assisted by compensatory hypertrophy of the columns of Bertin. Conversely, the morphogenesis of analgesic nephropathy explains why cessation of abuse in advanced renal disease has no significant effect on the progression of renal impairment. The residual renal parenchyma of the columns of Bertin is progressively damaged by overload hypertrophy and compensatory glomerular changes occur and loss of still functioning nephrons continues.

Differential diagnosis 

The early change, capillary sclerosis, is - so far as we know - specific to abuse of analgesics containing phenacetin. Other conditions associated with microangiopathy, e.g. diabetes mellitus and hypertension, do not produce changes which could be confused with capillary sclerosis. Capillary sclerosis also is unrelated to age, renal insufficiency or malignant tumours. Early papillary changes likewise are specific to analgesic nephropathy. Increase in interstitial ground substance and calcification also can occur with advancing age, but not in association with changes in peritubular capillaries or loops of Henle. Basement membrane thickening in the loops of Henle and peritubular capillaries is seen in diabetes mellitus and cirrhosis of the liver, but it does not lead to selective necrosis in the vicinity of the collecting duct bundle. It is hardly possible to mistake the non-reactive papillary necrosis in analgesic abuse with the papillary necrosis of acute pyelonephritis . The latter is yellow, friable and massively infiltrated with polymorphonuclear leukocytes.

The brown pigmentation of the urinary mucosa likewise is specific though it probably occurs only with abuse of phenacetin-containing analgesics. It is severe enough to be of diagnostic significance in only one third of cases.

Pathogenesis of “Analgesic nephropathy” due to phenacetin containing drugs

Capillary sclerosis of the urinary tract can serve as a model for understanding analgesic nephropathy, since anatomical and functional conditions in urinary mucosa are essentially less complex than in the kidney.

Of the factors considered to be involved in the pathogenesis of analgesic nephropathy - ischaemia, inhibition of prostaglandin synthesis, immunological factors and toxic damage - only toxic damage is believed to cause capillary sclerosis . The tree-ring morphology of the concentric layers of basement membrane suggests repeated toxic-induced endothelial necrosis. Each generation of cells develops its own basement membrane, if it does not lie directly on the old basement membrane, as is the case with capillary sclerosis caused by deposition of cell debris. The frequency and variable severity of capillary sclerosis in the different segments of the urinary tract is best explained by the fact that strongly lipophilic analgesic metabolites diffuse back out of urine and cause capillary damage. Thus capillary sclerosis is most marked at the points of maximum metabolite concentration and mucosal contact, i.e. distal renal pelvis, pelvi-ureteric junction and proximal ureter, as can readily be shown. The significant correlation between severity of capillary sclerosis on the one hand, and maximum daily dose of phenacetin on the other, emphasizes the importance of direct toxicity in the pathogenesis of capillary sclerosis. A noteworthy finding in this regard is that capillary sclerosis regresses in the absence of exposure to toxic analgesic metabolites. Ischaemia, inhibition of prostaglandin synthesis and immunological factors play no essential role in the pathogenesis of the morphological change of capillary sclerosis.

Similarly, toxicity is the common denominator of theories as to the pathogenesis of analgesic nephropathy. The topographical distribution of early changes in the inner medulla, with involvement of the loops of Henle, peritubular capillaries and interstitial cells, in the presence of well-preserved collecting ducts, supports the hypothesis that the concentration of the toxic agent increases with that of the urine and exerts a direct toxic effect at the points of maximum concentration. This assumption has been experimentally confirmed. Thus toxic compounds, e.g. paracetamol, aspirin and others, can reach concentrations in the renal medulla which are many times higher than in blood. This is especially true in the state of dehydration. The relationship between the total intake of phenacetin and the severity of the analgesic-induced renal lesion supports the hypothesis of toxic injury. Stabilization or improvement of renal function after cessation of abuse also is an argument in favour of the toxic origin of analgesic nephropathy. The toxic origin of phenacetin kidney is further supported by the biochemical evidence. It is probable that, under the influence of prostaglandin-endoperoxide synthase (PGES), highly reactive metabolites are formed from aspirin, paracetamol and phenacetin which then bind covalently to cell proteins, causing cell damage. The reactive metabolites are detoxified by binding to glutathione (GSH). The cells of the inner medulla have the highest concentrations of PGES and very low concentrations of GSH. Thus, excessive accumulation of highly reactive metabolites favour the toxic damage of tubular, endothelial and interstitial cells. Ischaemic necrosis is definitely not present, since the vasa recta running next to the necrotic papillae are unaffected. Furthermore, the selective occurrence of necrosis in the vicinity of the collecting duct bundle argues against the involvement of ischaemic necrosis. Immune processes likewise play no role in the development of papillary necrosis. In conclusion, removal of phenacetin from pain killers in Switzerland as in many other countries brought the solution: the “Phenacetin kidney disappeared from the daily clinical routine despite the fact that mixed analgesics have continued to be popular drugs, with paracetamol replacing phenacetin in many preparations since the early 1980s. So, phenacetin and its highly reactive metabolites are the major culprit for the “phenacetin kidney”.